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

Math Reasoning on MATH 500 (pass@4, Average)

87.8Success Rate (pass@4)

LLM-KD

Updated 1mo ago

Evaluation Results

Method	Links
LLM-KD 2026.04		87.8	63.54
JURY-RL 2026.04		86.6	64.04
Majority-Voting 2026.04		85.4	61.52
GT-Reward 2026.04		85.2	62.48
Entropy 2026.04		85.2	62.62
Self-Certainty 2026.04		84.6	60.33
CoReward 2026.04		83.8	59.99
LLM-KD 2026.04		82.2	53.51
LLM-as-a-Judge 2026.04		82.2	53.99
GT-Reward 2026.04		82	55.36
JURY-RL 2026.04		82	59.41
Entropy 2026.04		81.2	54.86
CoReward 2026.04		77.4	53.4
Self-Certainty 2026.04		77.2	55.17
Majority-Voting 2026.04		77.2	52.51
Before RL 2026.04		76.2	56.77
LLM-as-a-Judge 2026.04		74.6	50.35
Before RL 2026.04		72	49.59
Before RL 2026.04		65	46.26
LLM-KD 2026.04		65	46.21
JURY-RL 2026.04		64.8	48.48
GT-Reward 2026.04		63.8	45.46
Majority-Voting 2026.04		63.6	42.2
LLM-as-a-Judge 2026.04		62.2	45.93
CoReward 2026.04		60.4	43.51
Entropy 2026.04		58.8	42.77
Self-Certainty 2026.04		58	42.12