Files
GigaVerbo-v2-ablation-EDU-S…/evals.yaml
ModelHub XC dfd265b8c3 初始化项目,由ModelHub XC社区提供模型
Model: Polygl0t/GigaVerbo-v2-ablation-EDU-Synth-1.5B
Source: Original Platform
2026-09-20 15:44:19 +08:00

191 lines
9.6 KiB
YAML

evaluations:
arc_challenge_poly_pt_acc: 0.30085470085470084
arc_challenge_poly_pt_acc_norm: 0.3435897435897436
arc_challenge_poly_pt_acc_norm_stderr: 0.013889944781406437
arc_challenge_poly_pt_acc_stderr: 0.01341389388061822
arc_challenge_poly_pt_alias: arc_challenge_poly_pt
assin2_rte_acc,all: 0.5024509803921569
assin2_rte_acc_stderr,all: 0.007142229345039623
assin2_rte_alias: assin2_rte
assin2_rte_f1_macro,all: 0.34088811077510744
assin2_rte_f1_macro_stderr,all: 0.0036977314980645975
assin2_sts_alias: assin2_sts
assin2_sts_mse,all: 2.584869281045752
assin2_sts_mse_stderr,all: N/A
assin2_sts_pearson,all: 0.020246144461495013
assin2_sts_pearson_stderr,all: 0.012848574237206775
assin_entailment_acc: 0.62075
assin_entailment_acc_stderr: 0.007672651221656846
assin_entailment_alias: assin_entailment
assin_paraphrase_acc: 0.59675
assin_paraphrase_acc_stderr: 0.007757248423299025
assin_paraphrase_alias: assin_paraphrase
belebele_por_Latn_acc: 0.22666666666666666
belebele_por_Latn_acc_norm: 0.22666666666666666
belebele_por_Latn_acc_norm_stderr: 0.013963598349030474
belebele_por_Latn_acc_stderr: 0.013963598349030474
belebele_por_Latn_alias: belebele_por_Latn
bluex_acc,all: 0.2698191933240612
bluex_acc,exam_id__UNICAMP_2018: 0.3333333333333333
bluex_acc,exam_id__UNICAMP_2019: 0.3
bluex_acc,exam_id__UNICAMP_2020: 0.3090909090909091
bluex_acc,exam_id__UNICAMP_2021_1: 0.2391304347826087
bluex_acc,exam_id__UNICAMP_2021_2: 0.3333333333333333
bluex_acc,exam_id__UNICAMP_2022: 0.3076923076923077
bluex_acc,exam_id__UNICAMP_2023: 0.3023255813953488
bluex_acc,exam_id__UNICAMP_2024: 0.3333333333333333
bluex_acc,exam_id__USP_2018: 0.25925925925925924
bluex_acc,exam_id__USP_2019: 0.225
bluex_acc,exam_id__USP_2020: 0.17857142857142858
bluex_acc,exam_id__USP_2021: 0.25
bluex_acc,exam_id__USP_2022: 0.2653061224489796
bluex_acc,exam_id__USP_2023: 0.20454545454545456
bluex_acc,exam_id__USP_2024: 0.1951219512195122
bluex_acc_stderr,all: 0.009522448695577326
bluex_acc_stderr,exam_id__UNICAMP_2018: 0.03694964333964267
bluex_acc_stderr,exam_id__UNICAMP_2019: 0.03747275630361617
bluex_acc_stderr,exam_id__UNICAMP_2020: 0.035904653645853185
bluex_acc_stderr,exam_id__UNICAMP_2021_1: 0.036254833462246665
bluex_acc_stderr,exam_id__UNICAMP_2021_2: 0.038008168468976235
bluex_acc_stderr,exam_id__UNICAMP_2022: 0.042742868943786344
bluex_acc_stderr,exam_id__UNICAMP_2023: 0.040341644184605924
bluex_acc_stderr,exam_id__UNICAMP_2024: 0.04053327096189869
bluex_acc_stderr,exam_id__USP_2018: 0.0344262177569184
bluex_acc_stderr,exam_id__USP_2019: 0.038146963606575296
bluex_acc_stderr,exam_id__USP_2020: 0.029536187641256872
bluex_acc_stderr,exam_id__USP_2021: 0.03461792391082892
bluex_acc_stderr,exam_id__USP_2022: 0.036410693630710485
bluex_acc_stderr,exam_id__USP_2023: 0.03508480264125641
bluex_acc_stderr,exam_id__USP_2024: 0.03582337589286989
bluex_alias: bluex
calame_pt_acc: 0.5789980732177264
calame_pt_acc_stderr: 0.010838559109941895
calame_pt_alias: calame_pt
calame_pt_perplexity: 7.1322676051739915
calame_pt_perplexity_stderr: 0.41447318920967197
enem_challenge_acc,all: 0.198740377886634
enem_challenge_acc,exam_id__2009: 0.2
enem_challenge_acc,exam_id__2010: 0.21367521367521367
enem_challenge_acc,exam_id__2011: 0.2222222222222222
enem_challenge_acc,exam_id__2012: 0.2672413793103448
enem_challenge_acc,exam_id__2013: 0.21296296296296297
enem_challenge_acc,exam_id__2014: 0.1743119266055046
enem_challenge_acc,exam_id__2015: 0.16806722689075632
enem_challenge_acc,exam_id__2016: 0.18181818181818182
enem_challenge_acc,exam_id__2016_2: 0.2032520325203252
enem_challenge_acc,exam_id__2017: 0.21551724137931033
enem_challenge_acc,exam_id__2022: 0.19548872180451127
enem_challenge_acc,exam_id__2023: 0.14074074074074075
enem_challenge_acc_stderr,all: 0.00610558911077106
enem_challenge_acc_stderr,exam_id__2009: 0.02154694204401486
enem_challenge_acc_stderr,exam_id__2010: 0.021865515933679015
enem_challenge_acc_stderr,exam_id__2011: 0.022195336485659214
enem_challenge_acc_stderr,exam_id__2012: 0.02374992962172603
enem_challenge_acc_stderr,exam_id__2013: 0.02269408166705055
enem_challenge_acc_stderr,exam_id__2014: 0.020897673936382765
enem_challenge_acc_stderr,exam_id__2015: 0.019802918239207434
enem_challenge_acc_stderr,exam_id__2016: 0.020291042922884254
enem_challenge_acc_stderr,exam_id__2016_2: 0.020969856982989393
enem_challenge_acc_stderr,exam_id__2017: 0.02203067864108359
enem_challenge_acc_stderr,exam_id__2022: 0.019840331274268017
enem_challenge_acc_stderr,exam_id__2023: 0.017323886319225976
enem_challenge_alias: enem
faquad_nli_acc,all: 0.7753846153846153
faquad_nli_acc_stderr,all: 0.011564640936900579
faquad_nli_alias: faquad_nli
faquad_nli_f1_macro,all: 0.43674176776429807
faquad_nli_f1_macro_stderr,all: 0.0036705611888408394
global_piqa_completions_por_latn_braz_acc: 0.78
global_piqa_completions_por_latn_braz_acc_bytes: 0.76
global_piqa_completions_por_latn_braz_acc_bytes_stderr: 0.04292346959909278
global_piqa_completions_por_latn_braz_acc_norm: 0.75
global_piqa_completions_por_latn_braz_acc_norm_stderr: 0.04351941398892446
global_piqa_completions_por_latn_braz_acc_stderr: 0.041633319989322654
global_piqa_completions_por_latn_braz_alias: global_piqa_completions_por_latn_braz
hatebr_offensive_acc,all: 0.5457142857142857
hatebr_offensive_acc_stderr,all: 0.009408906567694409
hatebr_offensive_alias: hatebr_offensive_binary
hatebr_offensive_f1_macro,all: 0.4508029478016081
hatebr_offensive_f1_macro_stderr,all: 0.008913031329806937
hellaswag_poly_pt_acc: 0.361469281612309
hellaswag_poly_pt_acc_norm: 0.4601798678079965
hellaswag_poly_pt_acc_norm_stderr: 0.0051884131715642335
hellaswag_poly_pt_acc_stderr: 0.005001183649049966
hellaswag_poly_pt_alias: hellaswag_poly_pt
lambada_poly_pt_acc: 0.3904521637880846
lambada_poly_pt_acc_stderr: 0.0067967279472032245
lambada_poly_pt_alias: lambada_poly_pt
lambada_poly_pt_perplexity: 20.774403141688516
lambada_poly_pt_perplexity_stderr: 0.7231795341389268
mmlu_poly_pt_acc: 0.26208345842089464
mmlu_poly_pt_acc_stderr: 0.0038099775311724042
mmlu_poly_pt_alias: mmlu_poly_pt
oab_exams_acc,all: 0.2610478359908884
oab_exams_acc,exam_id__2010-01: 0.23529411764705882
oab_exams_acc,exam_id__2010-02: 0.27
oab_exams_acc,exam_id__2011-03: 0.2727272727272727
oab_exams_acc,exam_id__2011-04: 0.275
oab_exams_acc,exam_id__2011-05: 0.2625
oab_exams_acc,exam_id__2012-06: 0.2625
oab_exams_acc,exam_id__2012-06a: 0.225
oab_exams_acc,exam_id__2012-07: 0.275
oab_exams_acc,exam_id__2012-08: 0.25
oab_exams_acc,exam_id__2012-09: 0.35064935064935066
oab_exams_acc,exam_id__2013-10: 0.275
oab_exams_acc,exam_id__2013-11: 0.2875
oab_exams_acc,exam_id__2013-12: 0.25
oab_exams_acc,exam_id__2014-13: 0.25
oab_exams_acc,exam_id__2014-14: 0.275
oab_exams_acc,exam_id__2014-15: 0.32051282051282054
oab_exams_acc,exam_id__2015-16: 0.1875
oab_exams_acc,exam_id__2015-17: 0.2948717948717949
oab_exams_acc,exam_id__2015-18: 0.175
oab_exams_acc,exam_id__2016-19: 0.24358974358974358
oab_exams_acc,exam_id__2016-20: 0.25
oab_exams_acc,exam_id__2016-20a: 0.2875
oab_exams_acc,exam_id__2016-21: 0.2625
oab_exams_acc,exam_id__2017-22: 0.25
oab_exams_acc,exam_id__2017-23: 0.2125
oab_exams_acc,exam_id__2017-24: 0.275
oab_exams_acc,exam_id__2018-25: 0.275
oab_exams_acc_stderr,all: 0.005404522993204322
oab_exams_acc_stderr,exam_id__2010-01: 0.026605455313616685
oab_exams_acc_stderr,exam_id__2010-02: 0.025613225623037045
oab_exams_acc_stderr,exam_id__2011-03: 0.025797748277668113
oab_exams_acc_stderr,exam_id__2011-04: 0.028905026152293026
oab_exams_acc_stderr,exam_id__2011-05: 0.0283315547675685
oab_exams_acc_stderr,exam_id__2012-06: 0.028401435522187126
oab_exams_acc_stderr,exam_id__2012-06a: 0.02692493002179001
oab_exams_acc_stderr,exam_id__2012-07: 0.028685143633187495
oab_exams_acc_stderr,exam_id__2012-08: 0.02792646945667252
oab_exams_acc_stderr,exam_id__2012-09: 0.03143226713819032
oab_exams_acc_stderr,exam_id__2013-10: 0.028749520581305216
oab_exams_acc_stderr,exam_id__2013-11: 0.029216602612826752
oab_exams_acc_stderr,exam_id__2013-12: 0.02799652135475067
oab_exams_acc_stderr,exam_id__2014-13: 0.027941364058848093
oab_exams_acc_stderr,exam_id__2014-14: 0.028700665742904995
oab_exams_acc_stderr,exam_id__2014-15: 0.0304707537714559
oab_exams_acc_stderr,exam_id__2015-16: 0.025187032117302378
oab_exams_acc_stderr,exam_id__2015-17: 0.0297749047613837
oab_exams_acc_stderr,exam_id__2015-18: 0.02449033420128714
oab_exams_acc_stderr,exam_id__2016-19: 0.028105853302671797
oab_exams_acc_stderr,exam_id__2016-20: 0.02785679771062364
oab_exams_acc_stderr,exam_id__2016-20a: 0.029237922790613942
oab_exams_acc_stderr,exam_id__2016-21: 0.028334176853691478
oab_exams_acc_stderr,exam_id__2017-22: 0.027945788844504417
oab_exams_acc_stderr,exam_id__2017-23: 0.026388686173944978
oab_exams_acc_stderr,exam_id__2017-24: 0.02882934165027393
oab_exams_acc_stderr,exam_id__2018-25: 0.02893736358130262
oab_exams_alias: oab_exams
portuguese_hate_speech_acc,all: 0.700352526439483
portuguese_hate_speech_acc_stderr,all: 0.011075616750882897
portuguese_hate_speech_alias: portuguese_hate_speech_binary
portuguese_hate_speech_f1_macro,all: 0.4118866620594333
portuguese_hate_speech_f1_macro_stderr,all: 0.0038311957825553846
tweetsentbr_acc,all: 0.3373134328358209
tweetsentbr_acc_stderr,all: 0.0074549865355581415
tweetsentbr_alias: tweetsentbr
tweetsentbr_f1_macro,all: 0.21837645526737784
tweetsentbr_f1_macro_stderr,all: 0.005256512662126685
step: 22000