List your company

Comment Prediction on HealthBench (test)

7.31Medical Chat Annotation ScoreFeedback-to-Rubrics
-0.08441.83533.7555.6747May 28, 2026

Evaluation results

MethodLinks
Feedback-to-RubricsMethod Type=Rubric-based (Comment-wise refinement), Backbone=Gemini 3.1 Pro preview2026.057.31
Iterative rubric refinementMethod Type=Rubric-based (Ablation), Backbone=Gemini 3.1 Pro preview2026.057.15
Top-3 RAG with LLMMethod Type=Retrieval-based, Backbone=Gemini 3.1 Pro preview2026.057.08
Initial rubricMethod Type=Rubric-based (Ablation), Backbone=Gemini 3.1 Pro preview2026.056.72
No rubricMethod Type=Rubric-based (Ablation), Backbone=Gemini 3.1 Pro preview2026.056.62
Top-1 retrievalMethod Type=Retrieval-based, Backbone=Gemini 3.1 Pro preview2026.050.2