Share your thoughts, 1 month free Claude Pro on usSee more

Moral Reasoning on MoralExceptQA (challenge set)

64.47F1 Score

MORALCOT

Updated 3mo ago

Evaluation Results

Method	Links
MORALCOT 2022.10		64.47	66.05	66.96	0.38	3.2	62.1	70.68	54.04
Delphi++ 2022.10		58.27	62.16	76.79	0.34	1.34	36.61	70.6	40.81
InstructGPT 2022.10		53.94	64.36	98.52	0.38	1.59	42.4	70	50.48
BERT-large 2022.10		52.49	56.53	69.61	0.27	0.71	42.53	62.46	45.46
GPT3 2022.10		52.32	58.95	80.67	0.27	0.72	36.53	72.58	41.2
Random Baseline 2022.10		49.37	48.82	40.08	0.35	1	44.88	57.55	48.36
Delphi 2022.10		48.51	61.26	97.7	0.42	2.92	33.33	70.6	44.29
Always No 2022.10		45.99	60.81	100	0.258	0.7	33.33	70.6	33.33
BERT-base 2022.10		45.28	48.87	64.16	0.26	0.82	40.81	51.65	43.51
ROBERTa-large 2022.10		23.76	39.64	0.75	0.3	0.76	34.96	6.89	38.32
ALBERT-xxlarge 2022.10		22.07	39.19	0	0.46	1.41	33.33	6.89	33.33