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

Process Reward Modeling on PRM800K (test)

95.5Pos-Acc

ReasonEval-7B*

50.5262.197573.87585.5525Jun 8, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
95.595.521.221.260
2026.06
95.3695.3630.6630.6665.5
2026.06
83.2983.2950.8950.8964.3
2026.06
82.982.958.258.266.8
2026.06
52.2552.2573.9273.9250.5