phi-2-dpo-Nous.md

Model	AGIEval	GPT4All	TruthfulQA	Bigbench	Average
phi-2-dpo	30.39	71.68	50.75	34.9	46.93

AGIEval

Task	Version	Metric	Value		Stderr
agieval_aqua_rat	0	acc	24.02	±	2.69
		acc_norm	21.65	±	2.59
agieval_logiqa_en	0	acc	29.49	±	1.79
		acc_norm	31.34	±	1.82
agieval_lsat_ar	0	acc	19.57	±	2.62
		acc_norm	20.87	±	2.69
agieval_lsat_lr	0	acc	32.35	±	2.07
		acc_norm	26.67	±	1.96
agieval_lsat_rc	0	acc	36.06	±	2.93
		acc_norm	31.97	±	2.85
agieval_sat_en	0	acc	51.94	±	3.49
		acc_norm	47.57	±	3.49
agieval_sat_en_without_passage	0	acc	36.89	±	3.37
		acc_norm	33.98	±	3.31
agieval_sat_math	0	acc	31.82	±	3.15
		acc_norm	29.09	±	3.07

Average: 30.39%

GPT4All

Task	Version	Metric	Value		Stderr
arc_challenge	0	acc	52.56	±	1.46
		acc_norm	55.97	±	1.45
arc_easy	0	acc	81.10	±	0.80
		acc_norm	77.65	±	0.85
boolq	1	acc	82.39	±	0.67
hellaswag	0	acc	58.49	±	0.49
		acc_norm	76.55	±	0.42
openbookqa	0	acc	41.00	±	2.20
		acc_norm	51.60	±	2.24
piqa	0	acc	78.89	±	0.95
		acc_norm	79.82	±	0.94
winogrande	0	acc	77.82	±	1.17

Average: 71.68%

TruthfulQA

Task	Version	Metric	Value		Stderr
truthfulqa_mc	1	mc1	35.37	±	1.67
		mc2	50.75	±	1.55

Average: 50.75%

Bigbench

Task	Version	Metric	Value		Stderr
bigbench_causal_judgement	0	multiple_choice_grade	54.74	±	3.62
bigbench_date_understanding	0	multiple_choice_grade	57.45	±	2.58
bigbench_disambiguation_qa	0	multiple_choice_grade	39.15	±	3.04
bigbench_geometric_shapes	0	multiple_choice_grade	0.00	±	0.00
		exact_str_match	0.00	±	0.00
bigbench_logical_deduction_five_objects	0	multiple_choice_grade	24.80	±	1.93
bigbench_logical_deduction_seven_objects	0	multiple_choice_grade	16.29	±	1.40
bigbench_logical_deduction_three_objects	0	multiple_choice_grade	41.67	±	2.85
bigbench_movie_recommendation	0	multiple_choice_grade	35.00	±	2.14
bigbench_navigate	0	multiple_choice_grade	48.50	±	1.58
bigbench_reasoning_about_colored_objects	0	multiple_choice_grade	57.40	±	1.11
bigbench_ruin_names	0	multiple_choice_grade	25.45	±	2.06
bigbench_salient_translation_error_detection	0	multiple_choice_grade	25.35	±	1.38
bigbench_snarks	0	multiple_choice_grade	60.77	±	3.64
bigbench_sports_understanding	0	multiple_choice_grade	50.20	±	1.59
bigbench_temporal_sequences	0	multiple_choice_grade	18.30	±	1.22
bigbench_tracking_shuffled_objects_five_objects	0	multiple_choice_grade	18.64	±	1.10
bigbench_tracking_shuffled_objects_seven_objects	0	multiple_choice_grade	12.91	±	0.80
bigbench_tracking_shuffled_objects_three_objects	0	multiple_choice_grade	41.67	±	2.85

Average: 34.9%

Average score: 46.93% Elapsed time: 02:45:37