List your company

Question Answering on popQA Extreme

24.96F1 ScoreOpenDecoder
-0.59286.041112.67519.3089Jan 13, 2026

Evaluation results

MethodLinks
OpenDecoderBackbone=Qwen-2.5-3B-Instruct2026.0124.9622.02
RbFTBackbone=Qwen-2.5-3B-Instruct2026.0121.5920.8
Vanilla SFTBackbone=Qwen-2.5-3B-Instruct2026.0119.2718.37
AstuteRAGBackbone=Qwen-2.5-3B-Instruct2026.0115.7412.8
InstructRAGBackbone=Qwen-2.5-3B-Instruct2026.011.620.7
Vanilla RAGBackbone=Qwen-2.5-3B-Instruct2026.010.9811.87
RobustRAGBackbone=Qwen-2.5-3B-Instruct2026.010.391.2