List your company

Autoformalization on ProofNet (test)

44.09πFVMonotonic Reference-Free Refinement
28.989232.909636.8340.7504Jan 30, 2026

Evaluation results

MethodLinks
Monotonic Reference-Free RefinementJudges=GPT-4.1-mini, Time Step (t)=22026.0144.0932.9359.1447.5828.41——
Monotonic Reference-Free RefinementJudges=GPT-4.1-mini, Time Step (t)=32026.0144.0932.9358.9649.4628.97——
Monotonic Reference-Free RefinementJudges=GPT-4.1-mini, Time Step (t)=42026.0144.0933.3359.549.7329.38——
Monotonic Reference-Free RefinementJudges=GPT-4.1-mini, Time Step (t)=52026.0144.0933.8759.6850.2729.79——
Monotonic Reference-Free RefinementJudges=Qwen2.5-Coder-7B, Time Step (t)=22026.0144.0913.5829.218.878.44——
Monotonic Reference-Free RefinementJudges=Qwen2.5-Coder-7B, Time Step (t)=32026.0144.0912.928.858.68——
Monotonic Reference-Free RefinementJudges=Qwen2.5-Coder-7B, Time Step (t)=42026.0144.0912.929.038.337.97——
Monotonic Reference-Free RefinementJudges=Qwen2.5-Coder-7B, Time Step (t)=52026.0144.0912.929.218.338.03——
Monotonic Reference-Free RefinementJudges=GPT-4.1-mini, Time Step (t)=12026.0143.5530.7858.2445.1626.4——
Monotonic Reference-Free RefinementJudges=Qwen2.5-Coder-7B, Time Step (t)=12026.0143.5513.9829.398.338.15——
Monotonic Reference-Free RefinementJudges=GPT-4.1-mini, Time Step (t)=02026.0129.5726.3455.9137.6316.52——
Monotonic Reference-Free RefinementJudges=Qwen2.5-Coder-7B, Time Step (t)=02026.0129.5711.9629.397.85.18——
DRIFTRetrieval Strategy=DRIFT, Model=Claude-Opus-42025.10—————78.6119.79
Oracle*Retrieval Strategy=Oracle*, Model=Claude-Opus-42025.10—————81.2826.2
RAutoRetrieval Strategy=RAuto, Model=Claude-Opus-42025.10—————75.6719.25
Zero-shotRetrieval Strategy=Zero-shot, Model=Claude-Opus-42025.10—————68.4517.65