{ "results": { "arabicmmlu": { "acc,none": 0.4208232445520581, "acc_stderr,none": 0.004040113223189638, "alias": "arabicmmlu" }, "arabicmmlu_humanities": { "acc,none": 0.44239250275633957, "acc_stderr,none": 0.008046896182334524, "alias": " - Humanities" }, "arabicmmlu_high_history": { "alias": " - High History", "acc,none": 0.3144736842105263, "acc_stderr,none": 0.016853237146172328 }, "arabicmmlu_high_islamic_studies": { "alias": " - High Islamic Studies", "acc,none": 0.4221556886227545, "acc_stderr,none": 0.02706572265618471 }, "arabicmmlu_high_philosophy": { "alias": " - High Philosophy", "acc,none": 0.5128205128205128, "acc_stderr,none": 0.08108404256842 }, "arabicmmlu_islamic_studies": { "alias": " - Islamic Studies", "acc,none": 0.3489827856025039, "acc_stderr,none": 0.01887069517251757 }, "arabicmmlu_middle_history": { "alias": " - Middle History", "acc,none": 0.42857142857142855, "acc_stderr,none": 0.03481904844438804 }, "arabicmmlu_middle_islamic_studies": { "alias": " - Middle Islamic Studies", "acc,none": 0.49159663865546216, "acc_stderr,none": 0.03247390276569669 }, "arabicmmlu_primary_history": { "alias": " - Primary History", "acc,none": 0.37254901960784315, "acc_stderr,none": 0.04810840148082635 }, "arabicmmlu_primary_islamic_studies": { "alias": " - Primary Islamic Studies", "acc,none": 0.6016016016016016, "acc_stderr,none": 0.01549701356425835 }, "arabicmmlu_prof_law": { "alias": " - Prof Law", "acc,none": 0.4426751592356688, "acc_stderr,none": 0.028075313057827626 }, "arabicmmlu_language": { "acc,none": 0.4161603888213852, "acc_stderr,none": 0.011940274964070782, "alias": " - Language" }, "arabicmmlu_arabic_language_(general)": { "alias": " - Arabic Language (General)", "acc,none": 0.5098039215686274, "acc_stderr,none": 0.0202239460050743 }, "arabicmmlu_arabic_language_(grammar)": { "alias": " - Arabic Language (Grammar)", "acc,none": 0.3643835616438356, "acc_stderr,none": 0.02522471433569769 }, "arabicmmlu_high_arabic_language": { "alias": " - High Arabic Language", "acc,none": 0.27692307692307694, "acc_stderr,none": 0.022688042352424994 }, "arabicmmlu_middle_arabic_language": { "alias": " - Middle Arabic Language", "acc,none": 0.4444444444444444, "acc_stderr,none": 0.09745089103411436 }, "arabicmmlu_primary_arabic_language": { "alias": " - Primary Arabic Language", "acc,none": 0.47619047619047616, "acc_stderr,none": 0.031523917851640645 }, "arabicmmlu_other": { "acc,none": 0.47020933977455714, "acc_stderr,none": 0.009934531753088865, "alias": " - Other" }, "arabicmmlu_driving_test": { "alias": " - Driving Test", "acc,none": 0.5260115606936416, "acc_stderr,none": 0.014354525266560796 }, "arabicmmlu_general_knowledge": { "alias": " - General Knowledge", "acc,none": 0.3854166666666667, "acc_stderr,none": 0.016567242795987865 }, "arabicmmlu_middle_general_knowledge": { "alias": " - Middle General Knowledge", "acc,none": 0.45348837209302323, "acc_stderr,none": 0.03807016210250966 }, "arabicmmlu_primary_general_knowledge": { "alias": " - Primary General Knowledge", "acc,none": 0.4691358024691358, "acc_stderr,none": 0.03933037336475501 }, "arabicmmlu_univ_management": { "alias": " - Univ Management", "acc,none": 0.5866666666666667, "acc_stderr,none": 0.05724401171194134 }, "arabicmmlu_social_science": { "acc,none": 0.3818493150684932, "acc_stderr,none": 0.00812527639293321, "alias": " - Social Science" }, "arabicmmlu_high_civics": { "alias": " - High Civics", "acc,none": 0.28735632183908044, "acc_stderr,none": 0.048797477314965754 }, "arabicmmlu_high_economics": { "alias": " - High Economics", "acc,none": 0.425, "acc_stderr,none": 0.026090425569673732 }, "arabicmmlu_high_geography": { "alias": " - High Geography", "acc,none": 0.30346820809248554, "acc_stderr,none": 0.014277024139952538 }, "arabicmmlu_middle_civics": { "alias": " - Middle Civics", "acc,none": 0.3686440677966102, "acc_stderr,none": 0.031470730682346106 }, "arabicmmlu_middle_economics": { "alias": " - Middle Economics", "acc,none": 0.4827586206896552, "acc_stderr,none": 0.05388432214060092 }, "arabicmmlu_middle_geography": { "alias": " - Middle Geography", "acc,none": 0.3639705882352941, "acc_stderr,none": 0.029227192460032025 }, "arabicmmlu_middle_social_science": { "alias": " - Middle Social Science", "acc,none": 0.33195020746887965, "acc_stderr,none": 0.03039731808552683 }, "arabicmmlu_primary_geography": { "alias": " - Primary Geography", "acc,none": 0.43859649122807015, "acc_stderr,none": 0.0663095566682855 }, "arabicmmlu_primary_social_science": { "alias": " - Primary Social Science", "acc,none": 0.4978723404255319, "acc_stderr,none": 0.01884428842004545 }, "arabicmmlu_univ_accounting": { "alias": " - Univ Accounting", "acc,none": 0.4189189189189189, "acc_stderr,none": 0.05774600244608328 }, "arabicmmlu_univ_economics": { "alias": " - Univ Economics", "acc,none": 0.38686131386861317, "acc_stderr,none": 0.041762602685795874 }, "arabicmmlu_univ_political_science": { "alias": " - Univ Political Science", "acc,none": 0.36666666666666664, "acc_stderr,none": 0.03333333333333339 }, "arabicmmlu_stem": { "acc,none": 0.4030692139054181, "acc_stderr,none": 0.008590519358095423, "alias": " - STEM" }, "arabicmmlu_high_biology": { "alias": " - High Biology", "acc,none": 0.34776437189496096, "acc_stderr,none": 0.012692391957016312 }, "arabicmmlu_high_computer_science": { "alias": " - High Computer Science", "acc,none": 0.4099616858237548, "acc_stderr,none": 0.030501771826233565 }, "arabicmmlu_high_physics": { "alias": " - High Physics", "acc,none": 0.30196078431372547, "acc_stderr,none": 0.02880701939354399 }, "arabicmmlu_middle_computer_science": { "alias": " - Middle Computer Science", "acc,none": 0.5185185185185185, "acc_stderr,none": 0.09799078929868854 }, "arabicmmlu_middle_natural_science": { "alias": " - Middle Natural Science", "acc,none": 0.4256198347107438, "acc_stderr,none": 0.03184946380154992 }, "arabicmmlu_primary_computer_science": { "alias": " - Primary Computer Science", "acc,none": 0.46842105263157896, "acc_stderr,none": 0.03629703808831611 }, "arabicmmlu_primary_math": { "alias": " - Primary Math", "acc,none": 0.5476772616136919, "acc_stderr,none": 0.024640895323937397 }, "arabicmmlu_primary_natural_science": { "alias": " - Primary Natural Science", "acc,none": 0.46130952380952384, "acc_stderr,none": 0.02723600815931351 }, "arabicmmlu_univ_computer_science": { "alias": " - Univ Computer Science", "acc,none": 0.4375, "acc_stderr,none": 0.0625 } }, "groups": { "arabicmmlu": { "acc,none": 0.4208232445520581, "acc_stderr,none": 0.004040113223189638, "alias": "arabicmmlu" }, "arabicmmlu_humanities": { "acc,none": 0.44239250275633957, "acc_stderr,none": 0.008046896182334524, "alias": " - Humanities" }, "arabicmmlu_language": { "acc,none": 0.4161603888213852, "acc_stderr,none": 0.011940274964070782, "alias": " - Language" }, "arabicmmlu_other": { "acc,none": 0.47020933977455714, "acc_stderr,none": 0.009934531753088865, "alias": " - Other" }, "arabicmmlu_social_science": { "acc,none": 0.3818493150684932, "acc_stderr,none": 0.00812527639293321, "alias": " - Social Science" }, "arabicmmlu_stem": { "acc,none": 0.4030692139054181, "acc_stderr,none": 0.008590519358095423, "alias": " - STEM" } }, "group_subtasks": { "arabicmmlu_language": [ "arabicmmlu_arabic_language_(grammar)", "arabicmmlu_middle_arabic_language", "arabicmmlu_arabic_language_(general)", "arabicmmlu_primary_arabic_language", "arabicmmlu_high_arabic_language" ], "arabicmmlu_stem": [ "arabicmmlu_middle_computer_science", "arabicmmlu_primary_computer_science", "arabicmmlu_high_computer_science", "arabicmmlu_primary_natural_science", "arabicmmlu_middle_natural_science", "arabicmmlu_univ_computer_science", "arabicmmlu_high_physics", "arabicmmlu_high_biology", "arabicmmlu_primary_math" ], "arabicmmlu_humanities": [ "arabicmmlu_islamic_studies", "arabicmmlu_primary_history", "arabicmmlu_high_history", "arabicmmlu_primary_islamic_studies", "arabicmmlu_prof_law", "arabicmmlu_high_islamic_studies", "arabicmmlu_middle_islamic_studies", "arabicmmlu_high_philosophy", "arabicmmlu_middle_history" ], "arabicmmlu_social_science": [ "arabicmmlu_middle_economics", "arabicmmlu_univ_accounting", "arabicmmlu_high_geography", "arabicmmlu_univ_political_science", "arabicmmlu_middle_social_science", "arabicmmlu_univ_economics", "arabicmmlu_primary_geography", "arabicmmlu_middle_geography", "arabicmmlu_high_economics", "arabicmmlu_high_civics", "arabicmmlu_middle_civics", "arabicmmlu_primary_social_science" ], "arabicmmlu_other": [ "arabicmmlu_primary_general_knowledge", "arabicmmlu_driving_test", "arabicmmlu_univ_management", "arabicmmlu_middle_general_knowledge", "arabicmmlu_general_knowledge" ], "arabicmmlu": [ "arabicmmlu_other", "arabicmmlu_social_science", "arabicmmlu_humanities", "arabicmmlu_stem", "arabicmmlu_language" ] }, "configs": { "arabicmmlu_arabic_language_(general)": { "task": "arabicmmlu_arabic_language_(general)", "task_alias": "Arabic Language (General)", "tag": "arabicmmlu_language_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Arabic Language (General)", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_arabic_language_(grammar)": { "task": "arabicmmlu_arabic_language_(grammar)", "task_alias": "Arabic Language (Grammar)", "tag": "arabicmmlu_language_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Arabic Language (Grammar)", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_driving_test": { "task": "arabicmmlu_driving_test", "task_alias": "Driving Test", "tag": "arabicmmlu_other_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Driving Test", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_general_knowledge": { "task": "arabicmmlu_general_knowledge", "task_alias": "General Knowledge", "tag": "arabicmmlu_other_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "General Knowledge", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_high_arabic_language": { "task": "arabicmmlu_high_arabic_language", "task_alias": "High Arabic Language", "tag": "arabicmmlu_language_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "High Arabic Language", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_high_biology": { "task": "arabicmmlu_high_biology", "task_alias": "High Biology", "tag": "arabicmmlu_stem_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "High Biology", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_high_civics": { "task": "arabicmmlu_high_civics", "task_alias": "High Civics", "tag": "arabicmmlu_social_science_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "High Civics", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_high_computer_science": { "task": "arabicmmlu_high_computer_science", "task_alias": "High Computer Science", "tag": "arabicmmlu_stem_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "High Computer Science", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_high_economics": { "task": "arabicmmlu_high_economics", "task_alias": "High Economics", "tag": "arabicmmlu_social_science_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "High Economics", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_high_geography": { "task": "arabicmmlu_high_geography", "task_alias": "High Geography", "tag": "arabicmmlu_social_science_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "High Geography", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_high_history": { "task": "arabicmmlu_high_history", "task_alias": "High History", "tag": "arabicmmlu_humanities_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "High History", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_high_islamic_studies": { "task": "arabicmmlu_high_islamic_studies", "task_alias": "High Islamic Studies", "tag": "arabicmmlu_humanities_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "High Islamic Studies", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_high_philosophy": { "task": "arabicmmlu_high_philosophy", "task_alias": "High Philosophy", "tag": "arabicmmlu_humanities_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "High Philosophy", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_high_physics": { "task": "arabicmmlu_high_physics", "task_alias": "High Physics", "tag": "arabicmmlu_stem_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "High Physics", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_islamic_studies": { "task": "arabicmmlu_islamic_studies", "task_alias": "Islamic Studies", "tag": "arabicmmlu_humanities_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Islamic Studies", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_middle_arabic_language": { "task": "arabicmmlu_middle_arabic_language", "task_alias": "Middle Arabic Language", "tag": "arabicmmlu_language_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Middle Arabic Language", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_middle_civics": { "task": "arabicmmlu_middle_civics", "task_alias": "Middle Civics", "tag": "arabicmmlu_social_science_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Middle Civics", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_middle_computer_science": { "task": "arabicmmlu_middle_computer_science", "task_alias": "Middle Computer Science", "tag": "arabicmmlu_stem_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Middle Computer Science", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_middle_economics": { "task": "arabicmmlu_middle_economics", "task_alias": "Middle Economics", "tag": "arabicmmlu_social_science_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Middle Economics", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_middle_general_knowledge": { "task": "arabicmmlu_middle_general_knowledge", "task_alias": "Middle General Knowledge", "tag": "arabicmmlu_other_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Middle General Knowledge", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_middle_geography": { "task": "arabicmmlu_middle_geography", "task_alias": "Middle Geography", "tag": "arabicmmlu_social_science_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Middle Geography", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_middle_history": { "task": "arabicmmlu_middle_history", "task_alias": "Middle History", "tag": "arabicmmlu_humanities_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Middle History", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_middle_islamic_studies": { "task": "arabicmmlu_middle_islamic_studies", "task_alias": "Middle Islamic Studies", "tag": "arabicmmlu_humanities_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Middle Islamic Studies", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_middle_natural_science": { "task": "arabicmmlu_middle_natural_science", "task_alias": "Middle Natural Science", "tag": "arabicmmlu_stem_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Middle Natural Science", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_middle_social_science": { "task": "arabicmmlu_middle_social_science", "task_alias": "Middle Social Science", "tag": "arabicmmlu_social_science_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Middle Social Science", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_primary_arabic_language": { "task": "arabicmmlu_primary_arabic_language", "task_alias": "Primary Arabic Language", "tag": "arabicmmlu_language_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Primary Arabic Language", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_primary_computer_science": { "task": "arabicmmlu_primary_computer_science", "task_alias": "Primary Computer Science", "tag": "arabicmmlu_stem_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Primary Computer Science", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_primary_general_knowledge": { "task": "arabicmmlu_primary_general_knowledge", "task_alias": "Primary General Knowledge", "tag": "arabicmmlu_other_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Primary General Knowledge", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_primary_geography": { "task": "arabicmmlu_primary_geography", "task_alias": "Primary Geography", "tag": "arabicmmlu_social_science_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Primary Geography", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_primary_history": { "task": "arabicmmlu_primary_history", "task_alias": "Primary History", "tag": "arabicmmlu_humanities_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Primary History", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_primary_islamic_studies": { "task": "arabicmmlu_primary_islamic_studies", "task_alias": "Primary Islamic Studies", "tag": "arabicmmlu_humanities_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Primary Islamic Studies", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_primary_math": { "task": "arabicmmlu_primary_math", "task_alias": "Primary Math", "tag": "arabicmmlu_stem_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Primary Math", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_primary_natural_science": { "task": "arabicmmlu_primary_natural_science", "task_alias": "Primary Natural Science", "tag": "arabicmmlu_stem_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Primary Natural Science", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_primary_social_science": { "task": "arabicmmlu_primary_social_science", "task_alias": "Primary Social Science", "tag": "arabicmmlu_social_science_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Primary Social Science", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_prof_law": { "task": "arabicmmlu_prof_law", "task_alias": "Prof Law", "tag": "arabicmmlu_humanities_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Prof Law", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_univ_accounting": { "task": "arabicmmlu_univ_accounting", "task_alias": "Univ Accounting", "tag": "arabicmmlu_social_science_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Univ Accounting", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_univ_computer_science": { "task": "arabicmmlu_univ_computer_science", "task_alias": "Univ Computer Science", "tag": "arabicmmlu_stem_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Univ Computer Science", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_univ_economics": { "task": "arabicmmlu_univ_economics", "task_alias": "Univ Economics", "tag": "arabicmmlu_social_science_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Univ Economics", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_univ_management": { "task": "arabicmmlu_univ_management", "task_alias": "Univ Management", "tag": "arabicmmlu_other_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Univ Management", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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 } }, "arabicmmlu_univ_political_science": { "task": "arabicmmlu_univ_political_science", "task_alias": "Univ Political Science", "tag": "arabicmmlu_social_science_tasks", "dataset_path": "yazeed7/ArabicMMLU", "dataset_name": "Univ Political Science", "test_split": "test", "fewshot_split": "dev", "doc_to_text": "def doc_to_text(doc):\n \"\"\"\n Refactoring `prepare_data_en` to fit with the lm harness framework.\n https://github.com/mbzuai-nlp/ArabicMMLU/blob/main/util_prompt.py\n \"\"\"\n\n level = \"\" if not doc[\"Level\"] else \" for \" + level_en[doc[\"Level\"]]\n country = \"\" if not doc[\"Country\"] else \" in \" + doc[\"Country\"]\n main_meta_data = f\"{doc['Subject']} question{level}{country}\"\n\n question = (\n doc[\"Question\"]\n if doc[\"Context\"] == \"\"\n else f\"{doc['Context']}\\n\\n{doc['Question']}\"\n )\n\n options = []\n for i, opt in enumerate(\n [\"Option 1\", \"Option 2\", \"Option 3\", \"Option 4\", \"Option 5\"]\n ):\n if not doc[opt]:\n break\n options.append(f\"{alpa[i]} {doc[opt]}\")\n\n doc_text = PROMPT.format(main_meta_data, question, \"\\n\".join(options))\n\n return doc_text\n", "doc_to_target": "Answer Key", "doc_to_choice": "def doc_to_choice(doc):\n return [alpa[i][0] for i in range(5) if doc[f\"Option {i+1}\"]]\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "fewshot_config": { "sampler": "first_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": { "arabicmmlu": 0, "arabicmmlu_arabic_language_(general)": 0.0, "arabicmmlu_arabic_language_(grammar)": 0.0, "arabicmmlu_driving_test": 0.0, "arabicmmlu_general_knowledge": 0.0, "arabicmmlu_high_arabic_language": 0.0, "arabicmmlu_high_biology": 0.0, "arabicmmlu_high_civics": 0.0, "arabicmmlu_high_computer_science": 0.0, "arabicmmlu_high_economics": 0.0, "arabicmmlu_high_geography": 0.0, "arabicmmlu_high_history": 0.0, "arabicmmlu_high_islamic_studies": 0.0, "arabicmmlu_high_philosophy": 0.0, "arabicmmlu_high_physics": 0.0, "arabicmmlu_humanities": 0, "arabicmmlu_islamic_studies": 0.0, "arabicmmlu_language": 0, "arabicmmlu_middle_arabic_language": 0.0, "arabicmmlu_middle_civics": 0.0, "arabicmmlu_middle_computer_science": 0.0, "arabicmmlu_middle_economics": 0.0, "arabicmmlu_middle_general_knowledge": 0.0, "arabicmmlu_middle_geography": 0.0, "arabicmmlu_middle_history": 0.0, "arabicmmlu_middle_islamic_studies": 0.0, "arabicmmlu_middle_natural_science": 0.0, "arabicmmlu_middle_social_science": 0.0, "arabicmmlu_other": 0, "arabicmmlu_primary_arabic_language": 0.0, "arabicmmlu_primary_computer_science": 0.0, "arabicmmlu_primary_general_knowledge": 0.0, "arabicmmlu_primary_geography": 0.0, "arabicmmlu_primary_history": 0.0, "arabicmmlu_primary_islamic_studies": 0.0, "arabicmmlu_primary_math": 0.0, "arabicmmlu_primary_natural_science": 0.0, "arabicmmlu_primary_social_science": 0.0, "arabicmmlu_prof_law": 0.0, "arabicmmlu_social_science": 0, "arabicmmlu_stem": 0, "arabicmmlu_univ_accounting": 0.0, "arabicmmlu_univ_computer_science": 0.0, "arabicmmlu_univ_economics": 0.0, "arabicmmlu_univ_management": 0.0, "arabicmmlu_univ_political_science": 0.0 }, "n-shot": { "arabicmmlu_arabic_language_(general)": 0, "arabicmmlu_arabic_language_(grammar)": 0, "arabicmmlu_driving_test": 0, "arabicmmlu_general_knowledge": 0, "arabicmmlu_high_arabic_language": 0, "arabicmmlu_high_biology": 0, "arabicmmlu_high_civics": 0, "arabicmmlu_high_computer_science": 0, "arabicmmlu_high_economics": 0, "arabicmmlu_high_geography": 0, "arabicmmlu_high_history": 0, "arabicmmlu_high_islamic_studies": 0, "arabicmmlu_high_philosophy": 0, "arabicmmlu_high_physics": 0, "arabicmmlu_islamic_studies": 0, "arabicmmlu_middle_arabic_language": 0, "arabicmmlu_middle_civics": 0, "arabicmmlu_middle_computer_science": 0, "arabicmmlu_middle_economics": 0, "arabicmmlu_middle_general_knowledge": 0, "arabicmmlu_middle_geography": 0, "arabicmmlu_middle_history": 0, "arabicmmlu_middle_islamic_studies": 0, "arabicmmlu_middle_natural_science": 0, "arabicmmlu_middle_social_science": 0, "arabicmmlu_primary_arabic_language": 0, "arabicmmlu_primary_computer_science": 0, "arabicmmlu_primary_general_knowledge": 0, "arabicmmlu_primary_geography": 0, "arabicmmlu_primary_history": 0, "arabicmmlu_primary_islamic_studies": 0, "arabicmmlu_primary_math": 0, "arabicmmlu_primary_natural_science": 0, "arabicmmlu_primary_social_science": 0, "arabicmmlu_prof_law": 0, "arabicmmlu_univ_accounting": 0, "arabicmmlu_univ_computer_science": 0, "arabicmmlu_univ_economics": 0, "arabicmmlu_univ_management": 0, "arabicmmlu_univ_political_science": 0 }, "higher_is_better": { "arabicmmlu": { "acc": true }, "arabicmmlu_arabic_language_(general)": { "acc": true }, "arabicmmlu_arabic_language_(grammar)": { "acc": true }, "arabicmmlu_driving_test": { "acc": true }, "arabicmmlu_general_knowledge": { "acc": true }, "arabicmmlu_high_arabic_language": { "acc": true }, "arabicmmlu_high_biology": { "acc": true }, "arabicmmlu_high_civics": { "acc": true }, "arabicmmlu_high_computer_science": { "acc": true }, "arabicmmlu_high_economics": { "acc": true }, "arabicmmlu_high_geography": { "acc": true }, "arabicmmlu_high_history": { "acc": true }, "arabicmmlu_high_islamic_studies": { "acc": true }, "arabicmmlu_high_philosophy": { "acc": true }, "arabicmmlu_high_physics": { "acc": true }, "arabicmmlu_humanities": { "acc": true }, "arabicmmlu_islamic_studies": { "acc": true }, "arabicmmlu_language": { "acc": true }, "arabicmmlu_middle_arabic_language": { "acc": true }, "arabicmmlu_middle_civics": { "acc": true }, "arabicmmlu_middle_computer_science": { "acc": true }, "arabicmmlu_middle_economics": { "acc": true }, "arabicmmlu_middle_general_knowledge": { "acc": true }, "arabicmmlu_middle_geography": { "acc": true }, "arabicmmlu_middle_history": { "acc": true }, "arabicmmlu_middle_islamic_studies": { "acc": true }, "arabicmmlu_middle_natural_science": { "acc": true }, "arabicmmlu_middle_social_science": { "acc": true }, "arabicmmlu_other": { "acc": true }, "arabicmmlu_primary_arabic_language": { "acc": true }, "arabicmmlu_primary_computer_science": { "acc": true }, "arabicmmlu_primary_general_knowledge": { "acc": true }, "arabicmmlu_primary_geography": { "acc": true }, "arabicmmlu_primary_history": { "acc": true }, "arabicmmlu_primary_islamic_studies": { "acc": true }, "arabicmmlu_primary_math": { "acc": true }, "arabicmmlu_primary_natural_science": { "acc": true }, "arabicmmlu_primary_social_science": { "acc": true }, "arabicmmlu_prof_law": { "acc": true }, "arabicmmlu_social_science": { "acc": true }, "arabicmmlu_stem": { "acc": true }, "arabicmmlu_univ_accounting": { "acc": true }, "arabicmmlu_univ_computer_science": { "acc": true }, "arabicmmlu_univ_economics": { "acc": true }, "arabicmmlu_univ_management": { "acc": true }, "arabicmmlu_univ_political_science": { "acc": true } }, "n-samples": { "arabicmmlu_primary_general_knowledge": { "original": 162, "effective": 162 }, "arabicmmlu_driving_test": { "original": 1211, "effective": 1211 }, "arabicmmlu_univ_management": { "original": 75, "effective": 75 }, "arabicmmlu_middle_general_knowledge": { "original": 172, "effective": 172 }, "arabicmmlu_general_knowledge": { "original": 864, "effective": 864 }, "arabicmmlu_middle_economics": { "original": 87, "effective": 87 }, "arabicmmlu_univ_accounting": { "original": 74, "effective": 74 }, "arabicmmlu_high_geography": { "original": 1038, "effective": 1038 }, "arabicmmlu_univ_political_science": { "original": 210, "effective": 210 }, "arabicmmlu_middle_social_science": { "original": 241, "effective": 241 }, "arabicmmlu_univ_economics": { "original": 137, "effective": 137 }, "arabicmmlu_primary_geography": { "original": 57, "effective": 57 }, "arabicmmlu_middle_geography": { "original": 272, "effective": 272 }, "arabicmmlu_high_economics": { "original": 360, "effective": 360 }, "arabicmmlu_high_civics": { "original": 87, "effective": 87 }, "arabicmmlu_middle_civics": { "original": 236, "effective": 236 }, "arabicmmlu_primary_social_science": { "original": 705, "effective": 705 }, "arabicmmlu_islamic_studies": { "original": 639, "effective": 639 }, "arabicmmlu_primary_history": { "original": 102, "effective": 102 }, "arabicmmlu_high_history": { "original": 760, "effective": 760 }, "arabicmmlu_primary_islamic_studies": { "original": 999, "effective": 999 }, "arabicmmlu_prof_law": { "original": 314, "effective": 314 }, "arabicmmlu_high_islamic_studies": { "original": 334, "effective": 334 }, "arabicmmlu_middle_islamic_studies": { "original": 238, "effective": 238 }, "arabicmmlu_high_philosophy": { "original": 39, "effective": 39 }, "arabicmmlu_middle_history": { "original": 203, "effective": 203 }, "arabicmmlu_middle_computer_science": { "original": 27, "effective": 27 }, "arabicmmlu_primary_computer_science": { "original": 190, "effective": 190 }, "arabicmmlu_high_computer_science": { "original": 261, "effective": 261 }, "arabicmmlu_primary_natural_science": { "original": 336, "effective": 336 }, "arabicmmlu_middle_natural_science": { "original": 242, "effective": 242 }, "arabicmmlu_univ_computer_science": { "original": 64, "effective": 64 }, "arabicmmlu_high_physics": { "original": 255, "effective": 255 }, "arabicmmlu_high_biology": { "original": 1409, "effective": 1409 }, "arabicmmlu_primary_math": { "original": 409, "effective": 409 }, "arabicmmlu_arabic_language_(grammar)": { "original": 365, "effective": 365 }, "arabicmmlu_middle_arabic_language": { "original": 27, "effective": 27 }, "arabicmmlu_arabic_language_(general)": { "original": 612, "effective": 612 }, "arabicmmlu_primary_arabic_language": { "original": 252, "effective": 252 }, "arabicmmlu_high_arabic_language": { "original": 390, "effective": 390 } }, "config": { "model": "hf", "model_args": "pretrained=tiiuae/Falcon3-7B-Instruct,trust_remote_code=True,cache_dir=/tmp,parallelize=True", "model_num_parameters": 7455550464, "model_dtype": "torch.bfloat16", "model_revision": "main", "model_sha": "5563a370c1848366c7a095bde4bbff2cdb419cc6", "batch_size": 1, "batch_sizes": [], "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": "5e10e017", "date": 1736889500.3930833, "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-SXM4-80GB\nGPU 1: NVIDIA A100-SXM4-80GB\nGPU 2: NVIDIA A100-SXM4-80GB\nGPU 3: NVIDIA A100-SXM4-80GB\nGPU 4: NVIDIA A100-SXM4-80GB\nGPU 5: NVIDIA A100-SXM4-80GB\nGPU 6: NVIDIA A100-SXM4-80GB\nGPU 7: NVIDIA A100-SXM4-80GB\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): 96\nOn-line CPU(s) list: 0-95\nVendor ID: AuthenticAMD\nModel name: AMD EPYC 7V12 64-Core Processor\nCPU family: 23\nModel: 49\nThread(s) per core: 1\nCore(s) per socket: 48\nSocket(s): 2\nStepping: 0\nBogoMIPS: 4890.88\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 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 ssbd vmmcall fsgsbase bmi1 avx2 smep bmi2 rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 clzero xsaveerptr rdpru arat umip rdpid\nHypervisor vendor: Microsoft\nVirtualization type: full\nL1d cache: 3 MiB (96 instances)\nL1i cache: 3 MiB (96 instances)\nL2 cache: 48 MiB (96 instances)\nL3 cache: 384 MiB (24 instances)\nNUMA node(s): 4\nNUMA node0 CPU(s): 0-23\nNUMA node1 CPU(s): 24-47\nNUMA node2 CPU(s): 48-71\nNUMA node3 CPU(s): 72-95\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: Mitigation; untrained return thunk; SMT disabled\nVulnerability Spec rstack overflow: Mitigation; safe RET, no microcode\nVulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp\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==1.2.0\n[pip3] torchvision==0.19.0\n[pip3] triton==3.0.0\n[conda] Could not collect", "transformers_version": "4.48.0", "upper_git_hash": "f64fe2f2a86055aaecced603b56097fd79201711", "tokenizer_pad_token": [ "<|pad|>", "2023" ], "tokenizer_eos_token": [ "<|endoftext|>", "11" ], "tokenizer_bos_token": [ null, "None" ], "eot_token_id": 11, "max_length": 32768, "task_hashes": { "arabicmmlu_primary_general_knowledge": "91aa1e48a6f5ccff48fa6fa3277bbc97d23e6416fde69528f8956d0e90bc6244", "arabicmmlu_driving_test": "69f79faf8c303370c2df3ec536dd4c3cad19cf2cda6a1e77cff4852c0ebb14ee", "arabicmmlu_univ_management": "2ecfab399c12f6df05e9fd3a1db2573e7c48f5fa49566ce280a668a29896c4e3", "arabicmmlu_middle_general_knowledge": "e6929eb4f7ad78ad5b6b1141e390ce2c789a3ae9d3cf0ffeccac415a4212dcde", "arabicmmlu_general_knowledge": "1dfc3c92b60733bdc6f46f0f8268ac5feff7a327546595fff67ca2a4fa76ed4c", "arabicmmlu_middle_economics": "5583d598d2fe7413e6314b657d446ca334756888066f9fe8c2194c3a06458553", "arabicmmlu_univ_accounting": "9870f7d5ef58b1a884b890d26139fc3d9e3988082909e5b342eb220c40c74994", "arabicmmlu_high_geography": "d1b6c33005a743500852a60611f03f8827f80ab343290f5b1e5a3b1d4293e77f", "arabicmmlu_univ_political_science": "f27497dde305b538488920449e92ab0d4bfe35e4189e4212ad92e2fb76148e97", "arabicmmlu_middle_social_science": "1d25ab6b44a1b26de084ab850a913531b607a6e2803d8a79ce6863c832c22a95", "arabicmmlu_univ_economics": "f6cf4679eeae2e6eeb62050825cf38761c81d23b449aadd64e6adab85bbed352", "arabicmmlu_primary_geography": "e35f11ec9fef451aba1e9477b5d9486442f90bc2ea2e5e308a41c55dbce411fd", "arabicmmlu_middle_geography": "ed8b5cb8778ba57b3bfde2668f5c9bb71bff970583f294f428883c68bb9ae454", "arabicmmlu_high_economics": "a34142d316652408881b759c7330f4f661a7346e6771f5f22e85d19db23d7bfb", "arabicmmlu_high_civics": "a6da6b37a218224abfbdf5816c27d5c52546e3d4dbd6f7eed7a4979516c21acd", "arabicmmlu_middle_civics": "812fa9145e919b429ec0bda856075de404bf052193261c9dd9e07f80258b9b76", "arabicmmlu_primary_social_science": "58e86199fce5371c2af5e271fcf9beea7fb9947c6d72f921bc13d3caef2e7ec4", "arabicmmlu_islamic_studies": "6c27b44beb48e9774cb7d01f7b365291fe562fb35c8f2e0872f119b67f778c1c", "arabicmmlu_primary_history": "5f53990b8b0f0fe784c8297459f6591dbe8cbe04ce72de398525009c23591197", "arabicmmlu_high_history": "57c73385cc86d08a8d9da669118dd92f96b286567635901b01da5d79c898a4aa", "arabicmmlu_primary_islamic_studies": "f3d423a3b1b1b5b1128ea8428035df4b08c998c5450f38eb80cae4b79874fa2a", "arabicmmlu_prof_law": "731e89e57ca52310b4b446fea6ed38cebee09362abf58651c81281646b692f23", "arabicmmlu_high_islamic_studies": "a165da3444067e26499b01625e631eba032e28cc1fc6b6aa9030b53671452436", "arabicmmlu_middle_islamic_studies": "75c973380c08f25822af4104db06901b5c6a0cdb1a628d2bd90bcf8526a1ef5e", "arabicmmlu_high_philosophy": "8cbea21a7922a09751bd6d1eca16be8570a3544536dbe12de03731e194ad50c2", "arabicmmlu_middle_history": "1e1f90835724b6b5ce6297d91a656ac226e8210bdba020e4b9e3b6817a6414a9", "arabicmmlu_middle_computer_science": "249aebce1d740b259a5479569a981b9a343cd9fa8e309cfb0bcd53253c3a7a2e", "arabicmmlu_primary_computer_science": "27439beeb6cb7c0cc4cee804b7d1f9e7251a94c644aae58fd3206d35e2aa93d2", "arabicmmlu_high_computer_science": "a8c4e2ea301b4a23c47173ec3273d443028e21dc6fdb9d42e1b675220b4689ed", "arabicmmlu_primary_natural_science": "36e1bf3486dee3ffd262d051d99429869f8627e05fa7798d1d3f586992796fe7", "arabicmmlu_middle_natural_science": "a6a93f918a781ecab3b7ab692d645b199292403794c9bf2883ecb1ecace32e0b", "arabicmmlu_univ_computer_science": "5b38c4b463a4be775770bb51f341e7744c7e154afb4802e09cf199951621be99", "arabicmmlu_high_physics": "605d008475723d413ddeee9ab64db12fa85fa6ac0d0f029891694dfe5f7d3911", "arabicmmlu_high_biology": "04bfaf2a7d77c83199c24e03ab8f94c5e18b5006bc042901c3b6be100621a6db", "arabicmmlu_primary_math": "6169402e232f04147465bba4bc8be27e400675676c6d6c1951aabfadf2077e99", "arabicmmlu_arabic_language_(grammar)": "179324a1e8e0ccf6413411a6541c88130d78d31f0fec7274f9bfc19484f77a85", "arabicmmlu_middle_arabic_language": "7f510542f64580f95a35baf4533ed39fc59f6efe2a89af570675e4d9e30cf7f9", "arabicmmlu_arabic_language_(general)": "c342dca15e7dcbbe9b320f3726484abbec23656545fa3195a0014ad5d385e75c", "arabicmmlu_primary_arabic_language": "11ae5685e1cc66af215f4b43d45b2fcf6376e9389390c7e3aed3414122a935a1", "arabicmmlu_high_arabic_language": "2e5e0e90e40a42af3b2d5556d603782a252cc1350e65cf2654aaaa95e3e0cd06" }, "model_source": "hf", "model_name": "tiiuae/Falcon3-7B-Instruct", "model_name_sanitized": "tiiuae__Falcon3-7B-Instruct", "system_instruction": null, "system_instruction_sha": null, "fewshot_as_multiturn": false, "chat_template": null, "chat_template_sha": null, "start_time": 599750.782067174, "end_time": 599905.082863244, "total_evaluation_time_seconds": "154.30079607001971" }