List your company

General Knowledge Reasoning on MMLU Pro (pass@1)

70.93pass@1Critique-GRPO (Self-Critique & Self-Evaluation)
66.645267.757668.8769.9824Jun 3, 2025

Evaluation results

MethodLinks
Critique-GRPO (Self-Critique & Self-Evaluation)w/ External Supervision=false2025.0670.93
Critique-GRPO (Self-Critique)w/ External Supervision=true2025.0670.03
R1-GRPOw/ External Supervision=true2025.0670
Qwen3-8B (w/ Think)2025.0668.25
SFTw/ External Supervision=true2025.0666.81