List your company

Consistency Evaluation on DecisionBench

0.2902CR MeanDebate-AHP
-0.0048480.0717510.148350.224949Jan 23, 2026

Evaluation results

MethodLinks
Debate-AHPLLM generator=Llama3.1-8b2026.010.29020.43090
Standard-AHPLLM generator=Llama3.1-8b2026.010.16930.751471.42
Standard-AHPLLM generator=Llama3.1-70b2026.010.14141.392186.95
Doc2AHPLLM generator=Llama3.1-70b2026.010.0530.0957100
Standard-AHPLLM generator=GPT-5.22026.010.04340.316788.57
Doc2AHPLLM generator=GPT-5.22026.010.03260.0882100
Doc2AHPLLM generator=Llama3.1-8b2026.010.01390.0807100
Debate-AHPLLM generator=GPT-5.22026.010.01140.0335100
Debate-AHPLLM generator=Llama3.1-70b2026.010.00650.0132100