List your company

Infeasibility detection on minor-embedding

99.9Infeasibility Detection AccLlama-8b SFT
53.30865.40477.589.596Apr 1, 2026

Evaluation results

MethodLinks
Llama-8b SFTStrategy=BestOf82026.0499.910099.83.86
Llama-8b SFTStrategy=BestOf12026.0498.598992.3
GPT5.22026.0478.561.295.88.67
GPT5.2Reasoning=true2026.0477.55510049.68
Llama3.3-70B2026.047379673.72
Grok-4.1-FastReasoning=true2026.0471.242.4100114.4
Deepseek-V3.22026.0455.174.23617.37