2662 lines
131 KiB
JSON
2662 lines
131 KiB
JSON
|
{
|
||
|
|
"results": {
|
||
|
|
"openaimmlu": {
|
||
|
|
"acc,none": 0.44666001994017945,
|
||
|
|
"acc_stderr,none": 0.004112616445357971,
|
||
|
|
"alias": "openaimmlu"
|
||
|
|
},
|
||
|
|
"openaimmlu_STEM": {
|
||
|
|
"acc,none": 0.40794701986754967,
|
||
|
|
"acc_stderr,none": 0.008874683686325746,
|
||
|
|
"alias": " - STEM"
|
||
|
|
},
|
||
|
|
"openaimmlu_abstract_algebra": {
|
||
|
|
"alias": " - abstract_algebra",
|
||
|
|
"acc,none": 0.3,
|
||
|
|
"acc_stderr,none": 0.046056618647183814
|
||
|
|
},
|
||
|
|
"openaimmlu_astronomy": {
|
||
|
|
"alias": " - astronomy",
|
||
|
|
"acc,none": 0.5328947368421053,
|
||
|
|
"acc_stderr,none": 0.040601270352363966
|
||
|
|
},
|
||
|
|
"openaimmlu_college_biology": {
|
||
|
|
"alias": " - college_biology",
|
||
|
|
"acc,none": 0.4583333333333333,
|
||
|
|
"acc_stderr,none": 0.04166666666666665
|
||
|
|
},
|
||
|
|
"openaimmlu_college_chemistry": {
|
||
|
|
"alias": " - college_chemistry",
|
||
|
|
"acc,none": 0.43,
|
||
|
|
"acc_stderr,none": 0.04975698519562427
|
||
|
|
},
|
||
|
|
"openaimmlu_college_computer_science": {
|
||
|
|
"alias": " - college_computer_science",
|
||
|
|
"acc,none": 0.35,
|
||
|
|
"acc_stderr,none": 0.047937248544110196
|
||
|
|
},
|
||
|
|
"openaimmlu_college_mathematics": {
|
||
|
|
"alias": " - college_mathematics",
|
||
|
|
"acc,none": 0.35,
|
||
|
|
"acc_stderr,none": 0.0479372485441102
|
||
|
|
},
|
||
|
|
"openaimmlu_college_physics": {
|
||
|
|
"alias": " - college_physics",
|
||
|
|
"acc,none": 0.35294117647058826,
|
||
|
|
"acc_stderr,none": 0.04755129616062946
|
||
|
|
},
|
||
|
|
"openaimmlu_computer_security": {
|
||
|
|
"alias": " - computer_security",
|
||
|
|
"acc,none": 0.44,
|
||
|
|
"acc_stderr,none": 0.04988876515698589
|
||
|
|
},
|
||
|
|
"openaimmlu_conceptual_physics": {
|
||
|
|
"alias": " - conceptual_physics",
|
||
|
|
"acc,none": 0.37446808510638296,
|
||
|
|
"acc_stderr,none": 0.031639106653672915
|
||
|
|
},
|
||
|
|
"openaimmlu_econometrics": {
|
||
|
|
"alias": " - econometrics",
|
||
|
|
"acc,none": 0.2807017543859649,
|
||
|
|
"acc_stderr,none": 0.042270544512322
|
||
|
|
},
|
||
|
|
"openaimmlu_electrical_engineering": {
|
||
|
|
"alias": " - electrical_engineering",
|
||
|
|
"acc,none": 0.4413793103448276,
|
||
|
|
"acc_stderr,none": 0.04137931034482758
|
||
|
|
},
|
||
|
|
"openaimmlu_elementary_mathematics": {
|
||
|
|
"alias": " - elementary_mathematics",
|
||
|
|
"acc,none": 0.3783068783068783,
|
||
|
|
"acc_stderr,none": 0.024976954053155243
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_biology": {
|
||
|
|
"alias": " - high_school_biology",
|
||
|
|
"acc,none": 0.5419354838709678,
|
||
|
|
"acc_stderr,none": 0.028343787250540625
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_chemistry": {
|
||
|
|
"alias": " - high_school_chemistry",
|
||
|
|
"acc,none": 0.41379310344827586,
|
||
|
|
"acc_stderr,none": 0.03465304488406796
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_computer_science": {
|
||
|
|
"alias": " - high_school_computer_science",
|
||
|
|
"acc,none": 0.5,
|
||
|
|
"acc_stderr,none": 0.050251890762960605
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_mathematics": {
|
||
|
|
"alias": " - high_school_mathematics",
|
||
|
|
"acc,none": 0.35555555555555557,
|
||
|
|
"acc_stderr,none": 0.0291857149498574
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_physics": {
|
||
|
|
"alias": " - high_school_physics",
|
||
|
|
"acc,none": 0.3509933774834437,
|
||
|
|
"acc_stderr,none": 0.038969819642573754
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_statistics": {
|
||
|
|
"alias": " - high_school_statistics",
|
||
|
|
"acc,none": 0.3888888888888889,
|
||
|
|
"acc_stderr,none": 0.03324708911809117
|
||
|
|
},
|
||
|
|
"openaimmlu_humanities": {
|
||
|
|
"acc,none": 0.5144124168514412,
|
||
|
|
"acc_stderr,none": 0.011703005860087082,
|
||
|
|
"alias": " - Humanities"
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_european_history": {
|
||
|
|
"alias": " - high_school_european_history",
|
||
|
|
"acc,none": 0.5696969696969697,
|
||
|
|
"acc_stderr,none": 0.03866225962879077
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_us_history": {
|
||
|
|
"alias": " - high_school_us_history",
|
||
|
|
"acc,none": 0.5245098039215687,
|
||
|
|
"acc_stderr,none": 0.035050931943487976
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_world_history": {
|
||
|
|
"alias": " - high_school_world_history",
|
||
|
|
"acc,none": 0.5991561181434599,
|
||
|
|
"acc_stderr,none": 0.031900803894732356
|
||
|
|
},
|
||
|
|
"openaimmlu_international_law": {
|
||
|
|
"alias": " - international_law",
|
||
|
|
"acc,none": 0.6115702479338843,
|
||
|
|
"acc_stderr,none": 0.044492703500683836
|
||
|
|
},
|
||
|
|
"openaimmlu_jurisprudence": {
|
||
|
|
"alias": " - jurisprudence",
|
||
|
|
"acc,none": 0.5555555555555556,
|
||
|
|
"acc_stderr,none": 0.04803752235190192
|
||
|
|
},
|
||
|
|
"openaimmlu_logical_fallacies": {
|
||
|
|
"alias": " - logical_fallacies",
|
||
|
|
"acc,none": 0.4723926380368098,
|
||
|
|
"acc_stderr,none": 0.0392237829061099
|
||
|
|
},
|
||
|
|
"openaimmlu_philosophy": {
|
||
|
|
"alias": " - philosophy",
|
||
|
|
"acc,none": 0.47266881028938906,
|
||
|
|
"acc_stderr,none": 0.02835563356832818
|
||
|
|
},
|
||
|
|
"openaimmlu_prehistory": {
|
||
|
|
"alias": " - prehistory",
|
||
|
|
"acc,none": 0.4228395061728395,
|
||
|
|
"acc_stderr,none": 0.027487472980871598
|
||
|
|
},
|
||
|
|
"openaimmlu_world_religions": {
|
||
|
|
"alias": " - world_religions",
|
||
|
|
"acc,none": 0.5263157894736842,
|
||
|
|
"acc_stderr,none": 0.038295098689947286
|
||
|
|
},
|
||
|
|
"openaimmlu_other": {
|
||
|
|
"acc,none": 0.4364463924477411,
|
||
|
|
"acc_stderr,none": 0.00633626561036892,
|
||
|
|
"alias": " - Other"
|
||
|
|
},
|
||
|
|
"openaimmlu_anatomy": {
|
||
|
|
"alias": " - anatomy",
|
||
|
|
"acc,none": 0.37037037037037035,
|
||
|
|
"acc_stderr,none": 0.04171654161354544
|
||
|
|
},
|
||
|
|
"openaimmlu_clinical_knowledge": {
|
||
|
|
"alias": " - clinical_knowledge",
|
||
|
|
"acc,none": 0.5056603773584906,
|
||
|
|
"acc_stderr,none": 0.03077090076385131
|
||
|
|
},
|
||
|
|
"openaimmlu_college_medicine": {
|
||
|
|
"alias": " - college_medicine",
|
||
|
|
"acc,none": 0.4508670520231214,
|
||
|
|
"acc_stderr,none": 0.03794012674697029
|
||
|
|
},
|
||
|
|
"openaimmlu_formal_logic": {
|
||
|
|
"alias": " - formal_logic",
|
||
|
|
"acc,none": 0.3333333333333333,
|
||
|
|
"acc_stderr,none": 0.04216370213557835
|
||
|
|
},
|
||
|
|
"openaimmlu_global_facts": {
|
||
|
|
"alias": " - global_facts",
|
||
|
|
"acc,none": 0.34,
|
||
|
|
"acc_stderr,none": 0.04760952285695235
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_geography": {
|
||
|
|
"alias": " - high_school_geography",
|
||
|
|
"acc,none": 0.5858585858585859,
|
||
|
|
"acc_stderr,none": 0.035094383488796295
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_psychology": {
|
||
|
|
"alias": " - high_school_psychology",
|
||
|
|
"acc,none": 0.5431192660550459,
|
||
|
|
"acc_stderr,none": 0.021357458785226203
|
||
|
|
},
|
||
|
|
"openaimmlu_human_aging": {
|
||
|
|
"alias": " - human_aging",
|
||
|
|
"acc,none": 0.47533632286995514,
|
||
|
|
"acc_stderr,none": 0.03351695167652628
|
||
|
|
},
|
||
|
|
"openaimmlu_machine_learning": {
|
||
|
|
"alias": " - machine_learning",
|
||
|
|
"acc,none": 0.25,
|
||
|
|
"acc_stderr,none": 0.04109974682633932
|
||
|
|
},
|
||
|
|
"openaimmlu_medical_genetics": {
|
||
|
|
"alias": " - medical_genetics",
|
||
|
|
"acc,none": 0.56,
|
||
|
|
"acc_stderr,none": 0.04988876515698589
|
||
|
|
},
|
||
|
|
"openaimmlu_miscellaneous": {
|
||
|
|
"alias": " - miscellaneous",
|
||
|
|
"acc,none": 0.5440613026819924,
|
||
|
|
"acc_stderr,none": 0.01781040392543535
|
||
|
|
},
|
||
|
|
"openaimmlu_nutrition": {
|
||
|
|
"alias": " - nutrition",
|
||
|
|
"acc,none": 0.5294117647058824,
|
||
|
|
"acc_stderr,none": 0.028580341065138286
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_accounting": {
|
||
|
|
"alias": " - professional_accounting",
|
||
|
|
"acc,none": 0.3475177304964539,
|
||
|
|
"acc_stderr,none": 0.028406627809590947
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_law": {
|
||
|
|
"alias": " - professional_law",
|
||
|
|
"acc,none": 0.3396349413298566,
|
||
|
|
"acc_stderr,none": 0.01209559250693197
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_medicine": {
|
||
|
|
"alias": " - professional_medicine",
|
||
|
|
"acc,none": 0.47794117647058826,
|
||
|
|
"acc_stderr,none": 0.030343264224213528
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_psychology": {
|
||
|
|
"alias": " - professional_psychology",
|
||
|
|
"acc,none": 0.4035947712418301,
|
||
|
|
"acc_stderr,none": 0.019848280168401164
|
||
|
|
},
|
||
|
|
"openaimmlu_virology": {
|
||
|
|
"alias": " - virology",
|
||
|
|
"acc,none": 0.39156626506024095,
|
||
|
|
"acc_stderr,none": 0.03799857454479637
|
||
|
|
},
|
||
|
|
"openaimmlu_social_science": {
|
||
|
|
"acc,none": 0.46348143639683503,
|
||
|
|
"acc_stderr,none": 0.008379584468677955,
|
||
|
|
"alias": " - Social Science"
|
||
|
|
},
|
||
|
|
"openaimmlu_business_ethics": {
|
||
|
|
"alias": " - business_ethics",
|
||
|
|
"acc,none": 0.54,
|
||
|
|
"acc_stderr,none": 0.05009082659620332
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_government_and_politics": {
|
||
|
|
"alias": " - high_school_government_and_politics",
|
||
|
|
"acc,none": 0.5440414507772021,
|
||
|
|
"acc_stderr,none": 0.035944137112724366
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_macroeconomics": {
|
||
|
|
"alias": " - high_school_macroeconomics",
|
||
|
|
"acc,none": 0.46923076923076923,
|
||
|
|
"acc_stderr,none": 0.025302958890850154
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_microeconomics": {
|
||
|
|
"alias": " - high_school_microeconomics",
|
||
|
|
"acc,none": 0.5252100840336135,
|
||
|
|
"acc_stderr,none": 0.03243718055137411
|
||
|
|
},
|
||
|
|
"openaimmlu_human_sexuality": {
|
||
|
|
"alias": " - human_sexuality",
|
||
|
|
"acc,none": 0.5267175572519084,
|
||
|
|
"acc_stderr,none": 0.04379024936553894
|
||
|
|
},
|
||
|
|
"openaimmlu_management": {
|
||
|
|
"alias": " - management",
|
||
|
|
"acc,none": 0.5631067961165048,
|
||
|
|
"acc_stderr,none": 0.04911147107365777
|
||
|
|
},
|
||
|
|
"openaimmlu_marketing": {
|
||
|
|
"alias": " - marketing",
|
||
|
|
"acc,none": 0.6324786324786325,
|
||
|
|
"acc_stderr,none": 0.03158539157745636
|
||
|
|
},
|
||
|
|
"openaimmlu_moral_disputes": {
|
||
|
|
"alias": " - moral_disputes",
|
||
|
|
"acc,none": 0.47109826589595377,
|
||
|
|
"acc_stderr,none": 0.02687408588351835
|
||
|
|
},
|
||
|
|
"openaimmlu_moral_scenarios": {
|
||
|
|
"alias": " - moral_scenarios",
|
||
|
|
"acc,none": 0.2569832402234637,
|
||
|
|
"acc_stderr,none": 0.014614465821966342
|
||
|
|
},
|
||
|
|
"openaimmlu_public_relations": {
|
||
|
|
"alias": " - public_relations",
|
||
|
|
"acc,none": 0.4818181818181818,
|
||
|
|
"acc_stderr,none": 0.04785964010794916
|
||
|
|
},
|
||
|
|
"openaimmlu_security_studies": {
|
||
|
|
"alias": " - security_studies",
|
||
|
|
"acc,none": 0.5836734693877551,
|
||
|
|
"acc_stderr,none": 0.03155782816556164
|
||
|
|
},
|
||
|
|
"openaimmlu_sociology": {
|
||
|
|
"alias": " - sociology",
|
||
|
|
"acc,none": 0.6318407960199005,
|
||
|
|
"acc_stderr,none": 0.03410410565495302
|
||
|
|
},
|
||
|
|
"openaimmlu_us_foreign_policy": {
|
||
|
|
"alias": " - us_foreign_policy",
|
||
|
|
"acc,none": 0.65,
|
||
|
|
"acc_stderr,none": 0.047937248544110196
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"groups": {
|
||
|
|
"openaimmlu": {
|
||
|
|
"acc,none": 0.44666001994017945,
|
||
|
|
"acc_stderr,none": 0.004112616445357971,
|
||
|
|
"alias": "openaimmlu"
|
||
|
|
},
|
||
|
|
"openaimmlu_STEM": {
|
||
|
|
"acc,none": 0.40794701986754967,
|
||
|
|
"acc_stderr,none": 0.008874683686325746,
|
||
|
|
"alias": " - STEM"
|
||
|
|
},
|
||
|
|
"openaimmlu_humanities": {
|
||
|
|
"acc,none": 0.5144124168514412,
|
||
|
|
"acc_stderr,none": 0.011703005860087082,
|
||
|
|
"alias": " - Humanities"
|
||
|
|
},
|
||
|
|
"openaimmlu_other": {
|
||
|
|
"acc,none": 0.4364463924477411,
|
||
|
|
"acc_stderr,none": 0.00633626561036892,
|
||
|
|
"alias": " - Other"
|
||
|
|
},
|
||
|
|
"openaimmlu_social_science": {
|
||
|
|
"acc,none": 0.46348143639683503,
|
||
|
|
"acc_stderr,none": 0.008379584468677955,
|
||
|
|
"alias": " - Social Science"
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"group_subtasks": {
|
||
|
|
"openaimmlu_humanities": [
|
||
|
|
"openaimmlu_jurisprudence",
|
||
|
|
"openaimmlu_prehistory",
|
||
|
|
"openaimmlu_world_religions",
|
||
|
|
"openaimmlu_high_school_european_history",
|
||
|
|
"openaimmlu_logical_fallacies",
|
||
|
|
"openaimmlu_international_law",
|
||
|
|
"openaimmlu_high_school_us_history",
|
||
|
|
"openaimmlu_high_school_world_history",
|
||
|
|
"openaimmlu_philosophy"
|
||
|
|
],
|
||
|
|
"openaimmlu_social_science": [
|
||
|
|
"openaimmlu_high_school_government_and_politics",
|
||
|
|
"openaimmlu_human_sexuality",
|
||
|
|
"openaimmlu_high_school_microeconomics",
|
||
|
|
"openaimmlu_security_studies",
|
||
|
|
"openaimmlu_public_relations",
|
||
|
|
"openaimmlu_moral_disputes",
|
||
|
|
"openaimmlu_high_school_macroeconomics",
|
||
|
|
"openaimmlu_sociology",
|
||
|
|
"openaimmlu_marketing",
|
||
|
|
"openaimmlu_management",
|
||
|
|
"openaimmlu_business_ethics",
|
||
|
|
"openaimmlu_us_foreign_policy",
|
||
|
|
"openaimmlu_moral_scenarios"
|
||
|
|
],
|
||
|
|
"openaimmlu_other": [
|
||
|
|
"openaimmlu_nutrition",
|
||
|
|
"openaimmlu_professional_law",
|
||
|
|
"openaimmlu_clinical_knowledge",
|
||
|
|
"openaimmlu_college_medicine",
|
||
|
|
"openaimmlu_human_aging",
|
||
|
|
"openaimmlu_miscellaneous",
|
||
|
|
"openaimmlu_global_facts",
|
||
|
|
"openaimmlu_professional_medicine",
|
||
|
|
"openaimmlu_machine_learning",
|
||
|
|
"openaimmlu_professional_accounting",
|
||
|
|
"openaimmlu_high_school_psychology",
|
||
|
|
"openaimmlu_medical_genetics",
|
||
|
|
"openaimmlu_virology",
|
||
|
|
"openaimmlu_high_school_geography",
|
||
|
|
"openaimmlu_professional_psychology",
|
||
|
|
"openaimmlu_formal_logic",
|
||
|
|
"openaimmlu_anatomy"
|
||
|
|
],
|
||
|
|
"openaimmlu_STEM": [
|
||
|
|
"openaimmlu_high_school_mathematics",
|
||
|
|
"openaimmlu_college_computer_science",
|
||
|
|
"openaimmlu_college_chemistry",
|
||
|
|
"openaimmlu_high_school_chemistry",
|
||
|
|
"openaimmlu_econometrics",
|
||
|
|
"openaimmlu_astronomy",
|
||
|
|
"openaimmlu_college_physics",
|
||
|
|
"openaimmlu_computer_security",
|
||
|
|
"openaimmlu_high_school_statistics",
|
||
|
|
"openaimmlu_high_school_physics",
|
||
|
|
"openaimmlu_electrical_engineering",
|
||
|
|
"openaimmlu_elementary_mathematics",
|
||
|
|
"openaimmlu_high_school_computer_science",
|
||
|
|
"openaimmlu_abstract_algebra",
|
||
|
|
"openaimmlu_college_mathematics",
|
||
|
|
"openaimmlu_conceptual_physics",
|
||
|
|
"openaimmlu_high_school_biology",
|
||
|
|
"openaimmlu_college_biology"
|
||
|
|
],
|
||
|
|
"openaimmlu": [
|
||
|
|
"openaimmlu_STEM",
|
||
|
|
"openaimmlu_other",
|
||
|
|
"openaimmlu_social_science",
|
||
|
|
"openaimmlu_humanities"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"configs": {
|
||
|
|
"openaimmlu_abstract_algebra": {
|
||
|
|
"task": "openaimmlu_abstract_algebra",
|
||
|
|
"task_alias": "abstract_algebra",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "abstract_algebra",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_anatomy": {
|
||
|
|
"task": "openaimmlu_anatomy",
|
||
|
|
"task_alias": "anatomy",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "anatomy",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_astronomy": {
|
||
|
|
"task": "openaimmlu_astronomy",
|
||
|
|
"task_alias": "astronomy",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "astronomy",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_business_ethics": {
|
||
|
|
"task": "openaimmlu_business_ethics",
|
||
|
|
"task_alias": "business_ethics",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "business_ethics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_clinical_knowledge": {
|
||
|
|
"task": "openaimmlu_clinical_knowledge",
|
||
|
|
"task_alias": "clinical_knowledge",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "clinical_knowledge",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_college_biology": {
|
||
|
|
"task": "openaimmlu_college_biology",
|
||
|
|
"task_alias": "college_biology",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "college_biology",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_college_chemistry": {
|
||
|
|
"task": "openaimmlu_college_chemistry",
|
||
|
|
"task_alias": "college_chemistry",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "college_chemistry",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_college_computer_science": {
|
||
|
|
"task": "openaimmlu_college_computer_science",
|
||
|
|
"task_alias": "college_computer_science",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "college_computer_science",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_college_mathematics": {
|
||
|
|
"task": "openaimmlu_college_mathematics",
|
||
|
|
"task_alias": "college_mathematics",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "college_mathematics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_college_medicine": {
|
||
|
|
"task": "openaimmlu_college_medicine",
|
||
|
|
"task_alias": "college_medicine",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "college_medicine",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_college_physics": {
|
||
|
|
"task": "openaimmlu_college_physics",
|
||
|
|
"task_alias": "college_physics",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "college_physics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_computer_security": {
|
||
|
|
"task": "openaimmlu_computer_security",
|
||
|
|
"task_alias": "computer_security",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "computer_security",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_conceptual_physics": {
|
||
|
|
"task": "openaimmlu_conceptual_physics",
|
||
|
|
"task_alias": "conceptual_physics",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "conceptual_physics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_econometrics": {
|
||
|
|
"task": "openaimmlu_econometrics",
|
||
|
|
"task_alias": "econometrics",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "econometrics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_electrical_engineering": {
|
||
|
|
"task": "openaimmlu_electrical_engineering",
|
||
|
|
"task_alias": "electrical_engineering",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "electrical_engineering",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_elementary_mathematics": {
|
||
|
|
"task": "openaimmlu_elementary_mathematics",
|
||
|
|
"task_alias": "elementary_mathematics",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "elementary_mathematics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_formal_logic": {
|
||
|
|
"task": "openaimmlu_formal_logic",
|
||
|
|
"task_alias": "formal_logic",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "formal_logic",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_global_facts": {
|
||
|
|
"task": "openaimmlu_global_facts",
|
||
|
|
"task_alias": "global_facts",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "global_facts",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_biology": {
|
||
|
|
"task": "openaimmlu_high_school_biology",
|
||
|
|
"task_alias": "high_school_biology",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_biology",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_chemistry": {
|
||
|
|
"task": "openaimmlu_high_school_chemistry",
|
||
|
|
"task_alias": "high_school_chemistry",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_chemistry",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_computer_science": {
|
||
|
|
"task": "openaimmlu_high_school_computer_science",
|
||
|
|
"task_alias": "high_school_computer_science",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_computer_science",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_european_history": {
|
||
|
|
"task": "openaimmlu_high_school_european_history",
|
||
|
|
"task_alias": "high_school_european_history",
|
||
|
|
"tag": "openaimmlu_humanities_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_european_history",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_geography": {
|
||
|
|
"task": "openaimmlu_high_school_geography",
|
||
|
|
"task_alias": "high_school_geography",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_geography",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_government_and_politics": {
|
||
|
|
"task": "openaimmlu_high_school_government_and_politics",
|
||
|
|
"task_alias": "high_school_government_and_politics",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_government_and_politics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_macroeconomics": {
|
||
|
|
"task": "openaimmlu_high_school_macroeconomics",
|
||
|
|
"task_alias": "high_school_macroeconomics",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_macroeconomics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_mathematics": {
|
||
|
|
"task": "openaimmlu_high_school_mathematics",
|
||
|
|
"task_alias": "high_school_mathematics",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_mathematics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_microeconomics": {
|
||
|
|
"task": "openaimmlu_high_school_microeconomics",
|
||
|
|
"task_alias": "high_school_microeconomics",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_microeconomics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_physics": {
|
||
|
|
"task": "openaimmlu_high_school_physics",
|
||
|
|
"task_alias": "high_school_physics",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_physics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_psychology": {
|
||
|
|
"task": "openaimmlu_high_school_psychology",
|
||
|
|
"task_alias": "high_school_psychology",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_psychology",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_statistics": {
|
||
|
|
"task": "openaimmlu_high_school_statistics",
|
||
|
|
"task_alias": "high_school_statistics",
|
||
|
|
"tag": "openaimmlu_STEM_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_statistics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_us_history": {
|
||
|
|
"task": "openaimmlu_high_school_us_history",
|
||
|
|
"task_alias": "high_school_us_history",
|
||
|
|
"tag": "openaimmlu_humanities_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_us_history",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_world_history": {
|
||
|
|
"task": "openaimmlu_high_school_world_history",
|
||
|
|
"task_alias": "high_school_world_history",
|
||
|
|
"tag": "openaimmlu_humanities_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "high_school_world_history",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_human_aging": {
|
||
|
|
"task": "openaimmlu_human_aging",
|
||
|
|
"task_alias": "human_aging",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "human_aging",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_human_sexuality": {
|
||
|
|
"task": "openaimmlu_human_sexuality",
|
||
|
|
"task_alias": "human_sexuality",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "human_sexuality",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_international_law": {
|
||
|
|
"task": "openaimmlu_international_law",
|
||
|
|
"task_alias": "international_law",
|
||
|
|
"tag": "openaimmlu_humanities_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "international_law",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_jurisprudence": {
|
||
|
|
"task": "openaimmlu_jurisprudence",
|
||
|
|
"task_alias": "jurisprudence",
|
||
|
|
"tag": "openaimmlu_humanities_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "jurisprudence",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_logical_fallacies": {
|
||
|
|
"task": "openaimmlu_logical_fallacies",
|
||
|
|
"task_alias": "logical_fallacies",
|
||
|
|
"tag": "openaimmlu_humanities_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "logical_fallacies",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_machine_learning": {
|
||
|
|
"task": "openaimmlu_machine_learning",
|
||
|
|
"task_alias": "machine_learning",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "machine_learning",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_management": {
|
||
|
|
"task": "openaimmlu_management",
|
||
|
|
"task_alias": "management",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "management",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_marketing": {
|
||
|
|
"task": "openaimmlu_marketing",
|
||
|
|
"task_alias": "marketing",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "marketing",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_medical_genetics": {
|
||
|
|
"task": "openaimmlu_medical_genetics",
|
||
|
|
"task_alias": "medical_genetics",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "medical_genetics",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_miscellaneous": {
|
||
|
|
"task": "openaimmlu_miscellaneous",
|
||
|
|
"task_alias": "miscellaneous",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "miscellaneous",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_moral_disputes": {
|
||
|
|
"task": "openaimmlu_moral_disputes",
|
||
|
|
"task_alias": "moral_disputes",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "moral_disputes",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_moral_scenarios": {
|
||
|
|
"task": "openaimmlu_moral_scenarios",
|
||
|
|
"task_alias": "moral_scenarios",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "moral_scenarios",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_nutrition": {
|
||
|
|
"task": "openaimmlu_nutrition",
|
||
|
|
"task_alias": "nutrition",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "nutrition",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_philosophy": {
|
||
|
|
"task": "openaimmlu_philosophy",
|
||
|
|
"task_alias": "philosophy",
|
||
|
|
"tag": "openaimmlu_humanities_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "philosophy",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_prehistory": {
|
||
|
|
"task": "openaimmlu_prehistory",
|
||
|
|
"task_alias": "prehistory",
|
||
|
|
"tag": "openaimmlu_humanities_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "prehistory",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_accounting": {
|
||
|
|
"task": "openaimmlu_professional_accounting",
|
||
|
|
"task_alias": "professional_accounting",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "professional_accounting",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_law": {
|
||
|
|
"task": "openaimmlu_professional_law",
|
||
|
|
"task_alias": "professional_law",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "professional_law",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_medicine": {
|
||
|
|
"task": "openaimmlu_professional_medicine",
|
||
|
|
"task_alias": "professional_medicine",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "professional_medicine",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_psychology": {
|
||
|
|
"task": "openaimmlu_professional_psychology",
|
||
|
|
"task_alias": "professional_psychology",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "professional_psychology",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_public_relations": {
|
||
|
|
"task": "openaimmlu_public_relations",
|
||
|
|
"task_alias": "public_relations",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "public_relations",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_security_studies": {
|
||
|
|
"task": "openaimmlu_security_studies",
|
||
|
|
"task_alias": "security_studies",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "security_studies",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_sociology": {
|
||
|
|
"task": "openaimmlu_sociology",
|
||
|
|
"task_alias": "sociology",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "sociology",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_us_foreign_policy": {
|
||
|
|
"task": "openaimmlu_us_foreign_policy",
|
||
|
|
"task_alias": "us_foreign_policy",
|
||
|
|
"tag": "openaimmlu_social_science_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "us_foreign_policy",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_virology": {
|
||
|
|
"task": "openaimmlu_virology",
|
||
|
|
"task_alias": "virology",
|
||
|
|
"tag": "openaimmlu_other_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "virology",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"openaimmlu_world_religions": {
|
||
|
|
"task": "openaimmlu_world_religions",
|
||
|
|
"task_alias": "world_religions",
|
||
|
|
"tag": "openaimmlu_humanities_tasks",
|
||
|
|
"dataset_path": "khalidalt/openai_mmlu_arabic",
|
||
|
|
"dataset_name": "world_religions",
|
||
|
|
"test_split": "test",
|
||
|
|
"process_docs": "def process_docs(dataset: datasets.Dataset) -> datasets.Dataset:\n def _process_docs(doc):\n\n def format_example(doc, choices):\n options = []\n for _, choice in enumerate(choices):\n options.append(f'{en2ar[choice]}. {doc[choice]}')\n\n ar_subject = SUBJECTS[doc['Subject']]\n query = PROMPT.format(ar_subject, #doc['Subject'],\n doc['Question'],\n \"\\n\".join(options))\n return query\n\n keys_en = [\"A\", \"B\", \"C\", \"D\"]\n keys_ar = ['\u0623', '\u0628', '\u062c', '\u062f']\n ar_label = en2ar[doc['Answer']]\n\n out_doc = {\n \"query\": format_example(doc, keys_en),\n \"choices\": keys_ar,\n \"gold\": keys_ar.index(ar_label)\n }\n\n return out_doc\n\n return dataset.map(_process_docs) \n",
|
||
|
|
"doc_to_text": "query",
|
||
|
|
"doc_to_target": "gold",
|
||
|
|
"doc_to_choice": "choices",
|
||
|
|
"description": "",
|
||
|
|
"target_delimiter": " ",
|
||
|
|
"fewshot_delimiter": "\n\n",
|
||
|
|
"num_fewshot": 0,
|
||
|
|
"metric_list": [
|
||
|
|
{
|
||
|
|
"metric": "acc",
|
||
|
|
"aggregation": "mean",
|
||
|
|
"higher_is_better": true
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"output_type": "multiple_choice",
|
||
|
|
"repeats": 1,
|
||
|
|
"should_decontaminate": false,
|
||
|
|
"metadata": {
|
||
|
|
"version": 0.0
|
||
|
|
}
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"versions": {
|
||
|
|
"openaimmlu": 0,
|
||
|
|
"openaimmlu_STEM": 0,
|
||
|
|
"openaimmlu_abstract_algebra": 0.0,
|
||
|
|
"openaimmlu_anatomy": 0.0,
|
||
|
|
"openaimmlu_astronomy": 0.0,
|
||
|
|
"openaimmlu_business_ethics": 0.0,
|
||
|
|
"openaimmlu_clinical_knowledge": 0.0,
|
||
|
|
"openaimmlu_college_biology": 0.0,
|
||
|
|
"openaimmlu_college_chemistry": 0.0,
|
||
|
|
"openaimmlu_college_computer_science": 0.0,
|
||
|
|
"openaimmlu_college_mathematics": 0.0,
|
||
|
|
"openaimmlu_college_medicine": 0.0,
|
||
|
|
"openaimmlu_college_physics": 0.0,
|
||
|
|
"openaimmlu_computer_security": 0.0,
|
||
|
|
"openaimmlu_conceptual_physics": 0.0,
|
||
|
|
"openaimmlu_econometrics": 0.0,
|
||
|
|
"openaimmlu_electrical_engineering": 0.0,
|
||
|
|
"openaimmlu_elementary_mathematics": 0.0,
|
||
|
|
"openaimmlu_formal_logic": 0.0,
|
||
|
|
"openaimmlu_global_facts": 0.0,
|
||
|
|
"openaimmlu_high_school_biology": 0.0,
|
||
|
|
"openaimmlu_high_school_chemistry": 0.0,
|
||
|
|
"openaimmlu_high_school_computer_science": 0.0,
|
||
|
|
"openaimmlu_high_school_european_history": 0.0,
|
||
|
|
"openaimmlu_high_school_geography": 0.0,
|
||
|
|
"openaimmlu_high_school_government_and_politics": 0.0,
|
||
|
|
"openaimmlu_high_school_macroeconomics": 0.0,
|
||
|
|
"openaimmlu_high_school_mathematics": 0.0,
|
||
|
|
"openaimmlu_high_school_microeconomics": 0.0,
|
||
|
|
"openaimmlu_high_school_physics": 0.0,
|
||
|
|
"openaimmlu_high_school_psychology": 0.0,
|
||
|
|
"openaimmlu_high_school_statistics": 0.0,
|
||
|
|
"openaimmlu_high_school_us_history": 0.0,
|
||
|
|
"openaimmlu_high_school_world_history": 0.0,
|
||
|
|
"openaimmlu_human_aging": 0.0,
|
||
|
|
"openaimmlu_human_sexuality": 0.0,
|
||
|
|
"openaimmlu_humanities": 0,
|
||
|
|
"openaimmlu_international_law": 0.0,
|
||
|
|
"openaimmlu_jurisprudence": 0.0,
|
||
|
|
"openaimmlu_logical_fallacies": 0.0,
|
||
|
|
"openaimmlu_machine_learning": 0.0,
|
||
|
|
"openaimmlu_management": 0.0,
|
||
|
|
"openaimmlu_marketing": 0.0,
|
||
|
|
"openaimmlu_medical_genetics": 0.0,
|
||
|
|
"openaimmlu_miscellaneous": 0.0,
|
||
|
|
"openaimmlu_moral_disputes": 0.0,
|
||
|
|
"openaimmlu_moral_scenarios": 0.0,
|
||
|
|
"openaimmlu_nutrition": 0.0,
|
||
|
|
"openaimmlu_other": 0,
|
||
|
|
"openaimmlu_philosophy": 0.0,
|
||
|
|
"openaimmlu_prehistory": 0.0,
|
||
|
|
"openaimmlu_professional_accounting": 0.0,
|
||
|
|
"openaimmlu_professional_law": 0.0,
|
||
|
|
"openaimmlu_professional_medicine": 0.0,
|
||
|
|
"openaimmlu_professional_psychology": 0.0,
|
||
|
|
"openaimmlu_public_relations": 0.0,
|
||
|
|
"openaimmlu_security_studies": 0.0,
|
||
|
|
"openaimmlu_social_science": 0,
|
||
|
|
"openaimmlu_sociology": 0.0,
|
||
|
|
"openaimmlu_us_foreign_policy": 0.0,
|
||
|
|
"openaimmlu_virology": 0.0,
|
||
|
|
"openaimmlu_world_religions": 0.0
|
||
|
|
},
|
||
|
|
"n-shot": {
|
||
|
|
"openaimmlu_abstract_algebra": 0,
|
||
|
|
"openaimmlu_anatomy": 0,
|
||
|
|
"openaimmlu_astronomy": 0,
|
||
|
|
"openaimmlu_business_ethics": 0,
|
||
|
|
"openaimmlu_clinical_knowledge": 0,
|
||
|
|
"openaimmlu_college_biology": 0,
|
||
|
|
"openaimmlu_college_chemistry": 0,
|
||
|
|
"openaimmlu_college_computer_science": 0,
|
||
|
|
"openaimmlu_college_mathematics": 0,
|
||
|
|
"openaimmlu_college_medicine": 0,
|
||
|
|
"openaimmlu_college_physics": 0,
|
||
|
|
"openaimmlu_computer_security": 0,
|
||
|
|
"openaimmlu_conceptual_physics": 0,
|
||
|
|
"openaimmlu_econometrics": 0,
|
||
|
|
"openaimmlu_electrical_engineering": 0,
|
||
|
|
"openaimmlu_elementary_mathematics": 0,
|
||
|
|
"openaimmlu_formal_logic": 0,
|
||
|
|
"openaimmlu_global_facts": 0,
|
||
|
|
"openaimmlu_high_school_biology": 0,
|
||
|
|
"openaimmlu_high_school_chemistry": 0,
|
||
|
|
"openaimmlu_high_school_computer_science": 0,
|
||
|
|
"openaimmlu_high_school_european_history": 0,
|
||
|
|
"openaimmlu_high_school_geography": 0,
|
||
|
|
"openaimmlu_high_school_government_and_politics": 0,
|
||
|
|
"openaimmlu_high_school_macroeconomics": 0,
|
||
|
|
"openaimmlu_high_school_mathematics": 0,
|
||
|
|
"openaimmlu_high_school_microeconomics": 0,
|
||
|
|
"openaimmlu_high_school_physics": 0,
|
||
|
|
"openaimmlu_high_school_psychology": 0,
|
||
|
|
"openaimmlu_high_school_statistics": 0,
|
||
|
|
"openaimmlu_high_school_us_history": 0,
|
||
|
|
"openaimmlu_high_school_world_history": 0,
|
||
|
|
"openaimmlu_human_aging": 0,
|
||
|
|
"openaimmlu_human_sexuality": 0,
|
||
|
|
"openaimmlu_international_law": 0,
|
||
|
|
"openaimmlu_jurisprudence": 0,
|
||
|
|
"openaimmlu_logical_fallacies": 0,
|
||
|
|
"openaimmlu_machine_learning": 0,
|
||
|
|
"openaimmlu_management": 0,
|
||
|
|
"openaimmlu_marketing": 0,
|
||
|
|
"openaimmlu_medical_genetics": 0,
|
||
|
|
"openaimmlu_miscellaneous": 0,
|
||
|
|
"openaimmlu_moral_disputes": 0,
|
||
|
|
"openaimmlu_moral_scenarios": 0,
|
||
|
|
"openaimmlu_nutrition": 0,
|
||
|
|
"openaimmlu_philosophy": 0,
|
||
|
|
"openaimmlu_prehistory": 0,
|
||
|
|
"openaimmlu_professional_accounting": 0,
|
||
|
|
"openaimmlu_professional_law": 0,
|
||
|
|
"openaimmlu_professional_medicine": 0,
|
||
|
|
"openaimmlu_professional_psychology": 0,
|
||
|
|
"openaimmlu_public_relations": 0,
|
||
|
|
"openaimmlu_security_studies": 0,
|
||
|
|
"openaimmlu_sociology": 0,
|
||
|
|
"openaimmlu_us_foreign_policy": 0,
|
||
|
|
"openaimmlu_virology": 0,
|
||
|
|
"openaimmlu_world_religions": 0
|
||
|
|
},
|
||
|
|
"higher_is_better": {
|
||
|
|
"openaimmlu": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_STEM": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_abstract_algebra": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_anatomy": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_astronomy": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_business_ethics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_clinical_knowledge": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_college_biology": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_college_chemistry": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_college_computer_science": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_college_mathematics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_college_medicine": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_college_physics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_computer_security": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_conceptual_physics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_econometrics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_electrical_engineering": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_elementary_mathematics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_formal_logic": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_global_facts": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_biology": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_chemistry": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_computer_science": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_european_history": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_geography": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_government_and_politics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_macroeconomics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_mathematics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_microeconomics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_physics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_psychology": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_statistics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_us_history": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_world_history": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_human_aging": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_human_sexuality": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_humanities": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_international_law": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_jurisprudence": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_logical_fallacies": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_machine_learning": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_management": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_marketing": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_medical_genetics": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_miscellaneous": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_moral_disputes": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_moral_scenarios": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_nutrition": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_other": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_philosophy": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_prehistory": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_accounting": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_law": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_medicine": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_psychology": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_public_relations": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_security_studies": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_social_science": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_sociology": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_us_foreign_policy": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_virology": {
|
||
|
|
"acc": true
|
||
|
|
},
|
||
|
|
"openaimmlu_world_religions": {
|
||
|
|
"acc": true
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"n-samples": {
|
||
|
|
"openaimmlu_high_school_mathematics": {
|
||
|
|
"original": 270,
|
||
|
|
"effective": 270
|
||
|
|
},
|
||
|
|
"openaimmlu_college_computer_science": {
|
||
|
|
"original": 100,
|
||
|
|
"effective": 100
|
||
|
|
},
|
||
|
|
"openaimmlu_college_chemistry": {
|
||
|
|
"original": 100,
|
||
|
|
"effective": 100
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_chemistry": {
|
||
|
|
"original": 203,
|
||
|
|
"effective": 203
|
||
|
|
},
|
||
|
|
"openaimmlu_econometrics": {
|
||
|
|
"original": 114,
|
||
|
|
"effective": 114
|
||
|
|
},
|
||
|
|
"openaimmlu_astronomy": {
|
||
|
|
"original": 152,
|
||
|
|
"effective": 152
|
||
|
|
},
|
||
|
|
"openaimmlu_college_physics": {
|
||
|
|
"original": 102,
|
||
|
|
"effective": 102
|
||
|
|
},
|
||
|
|
"openaimmlu_computer_security": {
|
||
|
|
"original": 100,
|
||
|
|
"effective": 100
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_statistics": {
|
||
|
|
"original": 216,
|
||
|
|
"effective": 216
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_physics": {
|
||
|
|
"original": 151,
|
||
|
|
"effective": 151
|
||
|
|
},
|
||
|
|
"openaimmlu_electrical_engineering": {
|
||
|
|
"original": 145,
|
||
|
|
"effective": 145
|
||
|
|
},
|
||
|
|
"openaimmlu_elementary_mathematics": {
|
||
|
|
"original": 378,
|
||
|
|
"effective": 378
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_computer_science": {
|
||
|
|
"original": 100,
|
||
|
|
"effective": 100
|
||
|
|
},
|
||
|
|
"openaimmlu_abstract_algebra": {
|
||
|
|
"original": 100,
|
||
|
|
"effective": 100
|
||
|
|
},
|
||
|
|
"openaimmlu_college_mathematics": {
|
||
|
|
"original": 100,
|
||
|
|
"effective": 100
|
||
|
|
},
|
||
|
|
"openaimmlu_conceptual_physics": {
|
||
|
|
"original": 235,
|
||
|
|
"effective": 235
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_biology": {
|
||
|
|
"original": 310,
|
||
|
|
"effective": 310
|
||
|
|
},
|
||
|
|
"openaimmlu_college_biology": {
|
||
|
|
"original": 144,
|
||
|
|
"effective": 144
|
||
|
|
},
|
||
|
|
"openaimmlu_nutrition": {
|
||
|
|
"original": 306,
|
||
|
|
"effective": 306
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_law": {
|
||
|
|
"original": 1534,
|
||
|
|
"effective": 1534
|
||
|
|
},
|
||
|
|
"openaimmlu_clinical_knowledge": {
|
||
|
|
"original": 265,
|
||
|
|
"effective": 265
|
||
|
|
},
|
||
|
|
"openaimmlu_college_medicine": {
|
||
|
|
"original": 173,
|
||
|
|
"effective": 173
|
||
|
|
},
|
||
|
|
"openaimmlu_human_aging": {
|
||
|
|
"original": 223,
|
||
|
|
"effective": 223
|
||
|
|
},
|
||
|
|
"openaimmlu_miscellaneous": {
|
||
|
|
"original": 783,
|
||
|
|
"effective": 783
|
||
|
|
},
|
||
|
|
"openaimmlu_global_facts": {
|
||
|
|
"original": 100,
|
||
|
|
"effective": 100
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_medicine": {
|
||
|
|
"original": 272,
|
||
|
|
"effective": 272
|
||
|
|
},
|
||
|
|
"openaimmlu_machine_learning": {
|
||
|
|
"original": 112,
|
||
|
|
"effective": 112
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_accounting": {
|
||
|
|
"original": 282,
|
||
|
|
"effective": 282
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_psychology": {
|
||
|
|
"original": 545,
|
||
|
|
"effective": 545
|
||
|
|
},
|
||
|
|
"openaimmlu_medical_genetics": {
|
||
|
|
"original": 100,
|
||
|
|
"effective": 100
|
||
|
|
},
|
||
|
|
"openaimmlu_virology": {
|
||
|
|
"original": 166,
|
||
|
|
"effective": 166
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_geography": {
|
||
|
|
"original": 198,
|
||
|
|
"effective": 198
|
||
|
|
},
|
||
|
|
"openaimmlu_professional_psychology": {
|
||
|
|
"original": 612,
|
||
|
|
"effective": 612
|
||
|
|
},
|
||
|
|
"openaimmlu_formal_logic": {
|
||
|
|
"original": 126,
|
||
|
|
"effective": 126
|
||
|
|
},
|
||
|
|
"openaimmlu_anatomy": {
|
||
|
|
"original": 135,
|
||
|
|
"effective": 135
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_government_and_politics": {
|
||
|
|
"original": 193,
|
||
|
|
"effective": 193
|
||
|
|
},
|
||
|
|
"openaimmlu_human_sexuality": {
|
||
|
|
"original": 131,
|
||
|
|
"effective": 131
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_microeconomics": {
|
||
|
|
"original": 238,
|
||
|
|
"effective": 238
|
||
|
|
},
|
||
|
|
"openaimmlu_security_studies": {
|
||
|
|
"original": 245,
|
||
|
|
"effective": 245
|
||
|
|
},
|
||
|
|
"openaimmlu_public_relations": {
|
||
|
|
"original": 110,
|
||
|
|
"effective": 110
|
||
|
|
},
|
||
|
|
"openaimmlu_moral_disputes": {
|
||
|
|
"original": 346,
|
||
|
|
"effective": 346
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_macroeconomics": {
|
||
|
|
"original": 390,
|
||
|
|
"effective": 390
|
||
|
|
},
|
||
|
|
"openaimmlu_sociology": {
|
||
|
|
"original": 201,
|
||
|
|
"effective": 201
|
||
|
|
},
|
||
|
|
"openaimmlu_marketing": {
|
||
|
|
"original": 234,
|
||
|
|
"effective": 234
|
||
|
|
},
|
||
|
|
"openaimmlu_management": {
|
||
|
|
"original": 103,
|
||
|
|
"effective": 103
|
||
|
|
},
|
||
|
|
"openaimmlu_business_ethics": {
|
||
|
|
"original": 100,
|
||
|
|
"effective": 100
|
||
|
|
},
|
||
|
|
"openaimmlu_us_foreign_policy": {
|
||
|
|
"original": 100,
|
||
|
|
"effective": 100
|
||
|
|
},
|
||
|
|
"openaimmlu_moral_scenarios": {
|
||
|
|
"original": 895,
|
||
|
|
"effective": 895
|
||
|
|
},
|
||
|
|
"openaimmlu_jurisprudence": {
|
||
|
|
"original": 108,
|
||
|
|
"effective": 108
|
||
|
|
},
|
||
|
|
"openaimmlu_prehistory": {
|
||
|
|
"original": 324,
|
||
|
|
"effective": 324
|
||
|
|
},
|
||
|
|
"openaimmlu_world_religions": {
|
||
|
|
"original": 171,
|
||
|
|
"effective": 171
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_european_history": {
|
||
|
|
"original": 165,
|
||
|
|
"effective": 165
|
||
|
|
},
|
||
|
|
"openaimmlu_logical_fallacies": {
|
||
|
|
"original": 163,
|
||
|
|
"effective": 163
|
||
|
|
},
|
||
|
|
"openaimmlu_international_law": {
|
||
|
|
"original": 121,
|
||
|
|
"effective": 121
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_us_history": {
|
||
|
|
"original": 204,
|
||
|
|
"effective": 204
|
||
|
|
},
|
||
|
|
"openaimmlu_high_school_world_history": {
|
||
|
|
"original": 237,
|
||
|
|
"effective": 237
|
||
|
|
},
|
||
|
|
"openaimmlu_philosophy": {
|
||
|
|
"original": 311,
|
||
|
|
"effective": 311
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"config": {
|
||
|
|
"model": "hf",
|
||
|
|
"model_args": "pretrained=meta-llama/Meta-Llama-3.1-8B-Instruct,trust_remote_code=True,cache_dir=/tmp,parallelize=False",
|
||
|
|
"model_num_parameters": 8030261248,
|
||
|
|
"model_dtype": "torch.bfloat16",
|
||
|
|
"model_revision": "main",
|
||
|
|
"model_sha": "0e9e39f249a16976918f6564b8830bc894c89659",
|
||
|
|
"batch_size": "auto",
|
||
|
|
"batch_sizes": [
|
||
|
|
32
|
||
|
|
],
|
||
|
|
"device": null,
|
||
|
|
"use_cache": null,
|
||
|
|
"limit": null,
|
||
|
|
"bootstrap_iters": 100000,
|
||
|
|
"gen_kwargs": null,
|
||
|
|
"random_seed": 0,
|
||
|
|
"numpy_seed": 1234,
|
||
|
|
"torch_seed": 1234,
|
||
|
|
"fewshot_seed": 1234
|
||
|
|
},
|
||
|
|
"git_hash": "788a3672",
|
||
|
|
"date": 1737779004.899056,
|
||
|
|
"pretty_env_info": "PyTorch version: 2.4.0+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: Ubuntu 22.04.3 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0\nClang version: Could not collect\nCMake version: version 3.27.1\nLibc version: glibc-2.35\n\nPython version: 3.10.12 (main, Jun 11 2023, 05:26:28) [GCC 11.4.0] (64-bit runtime)\nPython platform: Linux-5.15.0-1064-azure-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.2.128\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: \nGPU 0: NVIDIA A100 80GB PCIe\nGPU 1: NVIDIA A100 80GB PCIe\n\nNvidia driver version: 535.161.08\ncuDNN version: Probably one of the following:\n/usr/lib/x86_64-linux-gnu/libcudnn.so.8.9.4\n/usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.9.4\n/usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.9.4\n/usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.9.4\n/usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.9.4\n/usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.9.4\n/usr/lib/x86_64-linux-gnu/libcudnn_ops_train.so.8.9.4\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture: x86_64\nCPU op-mode(s): 32-bit, 64-bit\nAddress sizes: 48 bits physical, 48 bits virtual\nByte Order: Little Endian\nCPU(s): 48\nOn-line CPU(s) list: 0-47\nVendor ID: AuthenticAMD\nModel name: AMD EPYC 7V13 64-Core Processor\nCPU family: 25\nModel: 1\nThread(s) per core: 1\nCore(s) per socket: 48\nSocket(s): 1\nStepping: 1\nBogoMIPS: 4890.87\nFlags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl tsc_reliable nonstop_tsc cpuid extd_apicid aperfmperf pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw topoext perfctr_core invpcid_single vmmcall fsgsbase bmi1 avx2 smep bmi2 erms invpcid rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves clzero xsaveerptr rdpru arat umip vaes vpclmulqdq rdpid fsrm\nHypervisor vendor: Microsoft\nVirtualization type: full\nL1d cache: 1.5 MiB (48 instances)\nL1i cache: 1.5 MiB (48 instances)\nL2 cache: 24 MiB (48 instances)\nL3 cache: 192 MiB (6 instances)\nNUMA node(s): 2\nNUMA node0 CPU(s): 0-23\nNUMA node1 CPU(s): 24-47\nVulnerability Gather data sampling: Not affected\nVulnerability Itlb multihit: Not affected\nVulnerability L1tf: Not affected\nVulnerability Mds: Not affected\nVulnerability Meltdown: Not affected\nVulnerability Mmio stale data: Not affected\nVulnerability Retbleed: Not affected\nVulnerability Spec rstack overflow: Mitigation; safe RET, no microcode\nVulnerability Spec store bypass: Vulnerable\nVulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2: Mitigation; Retpolines; STIBP disabled; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected\nVulnerability Srbds: Not affected\nVulnerability Tsx async abort: Not affected\n\nVersions of relevant libraries:\n[pip3] numpy==1.26.4\n[pip3] onnx==1.14.0\n[pip3] pytorch-lightning==2.0.7\n[pip3] pytorch-quantization==2.1.2\n[pip3] torch==2.4.0\n[pip3] torch-tensorrt==2.0.0.dev0\n[pip3] torchaudio==2.1.0\n[pip3] torchdata==0.7.0a0\n[pip3] torchmetrics==1.2.
|
||
|
|
"transformers_version": "4.48.1",
|
||
|
|
"upper_git_hash": "086919bd66f4e15fdcd4b792a7b27a698c1ba091",
|
||
|
|
"tokenizer_pad_token": [
|
||
|
|
"<|eot_id|>",
|
||
|
|
"128009"
|
||
|
|
],
|
||
|
|
"tokenizer_eos_token": [
|
||
|
|
"<|eot_id|>",
|
||
|
|
"128009"
|
||
|
|
],
|
||
|
|
"tokenizer_bos_token": [
|
||
|
|
"<|begin_of_text|>",
|
||
|
|
"128000"
|
||
|
|
],
|
||
|
|
"eot_token_id": 128009,
|
||
|
|
"max_length": 131072,
|
||
|
|
"task_hashes": {},
|
||
|
|
"model_source": "hf",
|
||
|
|
"model_name": "meta-llama/Meta-Llama-3.1-8B-Instruct",
|
||
|
|
"model_name_sanitized": "meta-llama__Meta-Llama-3.1-8B-Instruct",
|
||
|
|
"system_instruction": null,
|
||
|
|
"system_instruction_sha": null,
|
||
|
|
"fewshot_as_multiturn": false,
|
||
|
|
"chat_template": null,
|
||
|
|
"chat_template_sha": null,
|
||
|
|
"start_time": 26085.962482431,
|
||
|
|
"end_time": 26357.741487179,
|
||
|
|
"total_evaluation_time_seconds": "271.77900474799753"
|
||
|
|
}
|