List your company

Faithfulness Evaluation on Faithfulness Baseline Narratives

92.2Rank Accuracy (RA)Coherent Design
84.486.42588.4590.475Mar 20, 2026

Evaluation results

MethodLinks
Coherent DesignLLM=DeepSeek-V3.2-Exp2026.0392.29599.296.7
Critic-Rule DesignLLM=DeepSeek-V3.2-Exp2026.0391.29699.298.6
Critic DesignLLM=DeepSeek-V3.2-Exp2026.03919599.599
Critic-Rule DesignLLM=Mistral Medium 3.12026.0390.795.29998.9
Coherent-Rule DesignLLM=DeepSeek-V3.2-Exp2026.0390.79599.294.2
Basic DesignLLM=Claude Sonnet 4.52026.0390.595.79999.6
Basic DesignLLM=DeepSeek-V3.2-Exp2026.0390.59699.297.1
Critic DesignLLM=Claude Sonnet 4.52026.0390.5969998.3
Critic-Rule DesignLLM=Claude Sonnet 4.52026.0390.595.79998.5
Coherent DesignLLM=Claude Sonnet 4.52026.0390.595.79994.6
Coherent-Rule DesignLLM=Claude Sonnet 4.52026.0390.595.79995.3
Coherent-Rule DesignLLM=Mistral Medium 3.12026.0390.395.298.891.6
Critic DesignLLM=GPT-52026.0390.296.599.398.2
Basic DesignLLM=Mistral Medium 3.12026.0389.8959995.1
Basic DesignLLM=GPT-52026.0389.896.599.399
Critic DesignLLM=Mistral Medium 3.12026.0389.795.29998.8
Coherent DesignLLM=Mistral Medium 3.12026.0389.795.299.390.7
Coherent DesignLLM=GPT-52026.0389.396.399.395.4
Critic-Rule DesignLLM=GPT-52026.0388.596.399.398.9
Coherent-Rule DesignLLM=GPT-52026.0388.596.399.395.3
Critic DesignLLM=Llama 3.3 70B Instruct2026.0387.59599.396.8
Coherent DesignLLM=Llama 3.3 70B Instruct2026.0387.295.399.392.7
Coherent-Rule DesignLLM=Llama 3.3 70B Instruct2026.0386.595.299.389
Basic DesignLLM=Llama 3.3 70B Instruct2026.0385.79499.394.6
Critic-Rule DesignLLM=Llama 3.3 70B Instruct2026.0384.79499.397.3