Skip to content

Instantly share code, notes, and snippets.

@CultriX-Github
Created January 25, 2024 12:00
Show Gist options
  • Save CultriX-Github/b4f458fec63b5ece6cfe0057d830b21f to your computer and use it in GitHub Desktop.
Save CultriX-Github/b4f458fec63b5ece6cfe0057d830b21f to your computer and use it in GitHub Desktop.
Model AGIEval GPT4All TruthfulQA Bigbench Average
MergeTrix-7B 44.93 76.85 66.56 47.18 58.88

AGIEval

Task Version Metric Value Stderr
agieval_aqua_rat 0 acc 26.38 ± 2.77
acc_norm 25.20 ± 2.73
agieval_logiqa_en 0 acc 39.94 ± 1.92
acc_norm 40.09 ± 1.92
agieval_lsat_ar 0 acc 23.91 ± 2.82
acc_norm 20.87 ± 2.69
agieval_lsat_lr 0 acc 51.96 ± 2.21
acc_norm 52.35 ± 2.21
agieval_lsat_rc 0 acc 66.54 ± 2.88
acc_norm 66.17 ± 2.89
agieval_sat_en 0 acc 80.58 ± 2.76
acc_norm 79.61 ± 2.81
agieval_sat_en_without_passage 0 acc 43.20 ± 3.46
acc_norm 44.66 ± 3.47
agieval_sat_math 0 acc 32.27 ± 3.16
acc_norm 30.45 ± 3.11

Average: 44.93%

GPT4All

Task Version Metric Value Stderr
arc_challenge 0 acc 66.98 ± 1.37
acc_norm 67.15 ± 1.37
arc_easy 0 acc 86.91 ± 0.69
acc_norm 82.95 ± 0.77
boolq 1 acc 88.41 ± 0.56
hellaswag 0 acc 69.08 ± 0.46
acc_norm 86.01 ± 0.35
openbookqa 0 acc 37.20 ± 2.16
acc_norm 48.20 ± 2.24
piqa 0 acc 83.41 ± 0.87
acc_norm 84.66 ± 0.84
winogrande 0 acc 80.58 ± 1.11

Average: 76.85%

TruthfulQA

Task Version Metric Value Stderr
truthfulqa_mc 1 mc1 51.04 ± 1.75
mc2 66.56 ± 1.53

Average: 66.56%

Bigbench

Task Version Metric Value Stderr
bigbench_causal_judgement 0 multiple_choice_grade 57.89 ± 3.59
bigbench_date_understanding 0 multiple_choice_grade 65.58 ± 2.48
bigbench_disambiguation_qa 0 multiple_choice_grade 40.31 ± 3.06
bigbench_geometric_shapes 0 multiple_choice_grade 24.79 ± 2.28
exact_str_match 0.00 ± 0.00
bigbench_logical_deduction_five_objects 0 multiple_choice_grade 35.00 ± 2.14
bigbench_logical_deduction_seven_objects 0 multiple_choice_grade 24.29 ± 1.62
bigbench_logical_deduction_three_objects 0 multiple_choice_grade 58.00 ± 2.85
bigbench_movie_recommendation 0 multiple_choice_grade 43.60 ± 2.22
bigbench_navigate 0 multiple_choice_grade 55.40 ± 1.57
bigbench_reasoning_about_colored_objects 0 multiple_choice_grade 72.90 ± 0.99
bigbench_ruin_names 0 multiple_choice_grade 52.68 ± 2.36
bigbench_salient_translation_error_detection 0 multiple_choice_grade 31.16 ± 1.47
bigbench_snarks 0 multiple_choice_grade 71.27 ± 3.37
bigbench_sports_understanding 0 multiple_choice_grade 75.56 ± 1.37
bigbench_temporal_sequences 0 multiple_choice_grade 41.70 ± 1.56
bigbench_tracking_shuffled_objects_five_objects 0 multiple_choice_grade 22.80 ± 1.19
bigbench_tracking_shuffled_objects_seven_objects 0 multiple_choice_grade 18.23 ± 0.92
bigbench_tracking_shuffled_objects_three_objects 0 multiple_choice_grade 58.00 ± 2.85

Average: 47.18%

Average score: 58.88%

Elapsed time: 02:12:36

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment