Question Answering on MMLU Law Large Shift (test) benchmark leaderboard · SOTA2 Research

Question Answering on MMLU Law Large Shift (test)

37.9AccuracyDropout

Evaluation Results

Method	Links
Dropout2026.01		37.9	34.2	2.09
Bayesian-LoRA2026.01		37.9	16.5	1.42
Ckpt Ens2026.01		37.7	32.1	1.94
MAP2026.01		37.4	35.2	2.12
Temp2026.01		37.4	22.7	1.65
BLoB2026.01		37.4	25.3	2.01
LA2026.01		37.3	23.9	1.68
LLLA2026.01		37.1	33.6	1.49
BBB2026.01		36.1	28.1	1.65