List your company

Refusal on Irrelevant Queries on LongBench

97.52WikiMQA ScoreDoc2Atom
3.928.252.576.8Jun 10, 2026

Evaluation results

MethodLinks
Doc2AtomBase Model=Gemma-2-2B-It2026.0697.590.1795.2295.9595.3384.3361.564
D2LatomBase Model=Gemma-2-2B-It2026.062215.8311.3311.512.524.522.335.5
Doc2AtomBase Model=Qwen3-4B-Instruct2026.0617.6419.08649528.5892.593.2520
D2LrawBase Model=Gemma-2-2B-It2026.0615.83161840.58.5721.158.5
D2LckptBase Model=Qwen3-4B-Instruct2026.0613.215.2817.9621.4714.3875.0438.356.54
D2LrawBase Model=Qwen3-4B-Instruct2026.061110.517.33336.55.550.54
D2LckptBase Model=Gemma-2-2B-It2026.068.582.6714.5311.51
D2LatomBase Model=Qwen3-4B-Instruct2026.067.510.520.67216.5740.333