List your company

Comment Prediction on ExpertLongBench (test)

2.75Bio ScoreFeedback-to-Rubrics
0.47241.06371.6552.2463May 28, 2026

Evaluation results

MethodLinks
Feedback-to-RubricsMethod Type=Rubric-based (Comment-wise refinement), Backbone=Gemini 3.1 Pro preview2026.052.753.737.433.954.752.31
Initial rubricMethod Type=Rubric-based (Ablation), Backbone=Gemini 3.1 Pro preview2026.052.663.427.133.963.842.43
Iterative rubric refinementMethod Type=Rubric-based (Ablation), Backbone=Gemini 3.1 Pro preview2026.052.613.6474.114.212.35
Top-3 RAG with LLMMethod Type=Retrieval-based, Backbone=Gemini 3.1 Pro preview2026.052.33.124.762.262.752.06
Top-1 retrievalMethod Type=Retrieval-based, Backbone=Gemini 3.1 Pro preview2026.051.681.563.342.051.390.28
No rubricMethod Type=Rubric-based (Ablation), Backbone=Gemini 3.1 Pro preview2026.050.561.761.181.371.821.79