Natural Language Understanding and Generation on PromptSource T0 (held-out) benchmark leaderboard · SOTA2 Research

Natural Language Understanding and Generation on PromptSource T0 (held-out)

44.3ANLI R1OPT-IML 175B

Evaluation Results

Method	Links
OPT-IML 175B2022.12		44.3	39.9	43.5	75.6	82.7	93.3	53.8	53.8	56.9	16.4	72.7	34.4	55.6
OPT-IML 175B2022.12		42.2	38.5	39.6	56.4	73.4	95	59.2	53.6	56.6	16.3	72.7	34.4	53.2
T0-original-task 11B2022.12		42.1	37.9	39.7	58.5	80.2	96.7	58.6	56	62.5	6.2	83.8	24	53.8
OPT-IML 30B2022.12		38.3	35	38.8	66.8	65.1	85.6	62.4	54.2	52.9	15.9	64.6	22.3	50.1
OPT 1752022.12		37.8	34.7	36.5	63.5	51.6	63.5	40.2	49.9	52.2	16.1	53.9	85.5	48.8
OPT-IML 30B2022.12		37.1	35.4	36.6	43.2	67.8	90.7	58.2	54.7	53.4	15.7	64.6	22.3	48.3
OPT 1752022.12		34.1	34.1	34.7	38.9	54	57	51	49.7	50.1	15	53.9	85.5	46.5
OPT 302022.12		33.7	34.1	34.7	24.6	56.4	55.5	43.5	50.8	50.2	15.2	54.9	85.5	44.9
OPT 302022.12		33.6	33.2	33.3	43.6	49.6	55.7	45.5	50.7	50.2	15.7	54.9	85.5	45.9
T0-original-task 11B2022.12		33.6	33.1	33.2	48.9	47.3	94.1	63.5	50	57.9	4.5	83.8	24	47.8