List your company

Formal theorem proving on Lean (test)

73Pass@1α-DPG
53.44858.52463.668.676Dec 5, 2025

Evaluation results

MethodLinks
α-DPGalpha=0.9992025.127383.5——
RLOO2025.1271.883.5——
GPG2025.1271.680——
GRPO2025.1271.581——
ReMax2025.1270.485——
GRPO-Rw-Ulkly2025.1267.486——
α-DPGalpha=0.92025.1266.286——
α-DPGalpha=0.752025.1262.687——
α-DPGalpha=0.52025.1261.488.5——
GRPOKL constraint=High2025.1259.886.5——
α-DPGalpha=0.252025.1259.887——
α-DPGalpha=02025.1258.988——
GRPO-Pass@k2025.1254.887——
Base SFT2025.1254.286——