List your company

Question Answering on 2Wiki Extreme

26.99F1 ScoreOpenDecoder
0.14767.116314.08521.0537Jan 13, 2026

Evaluation results

MethodLinks
OpenDecoderBackbone=Qwen-2.5-3B-Instruct2026.0126.9924
RbFTBackbone=Qwen-2.5-3B-Instruct2026.0124.2822.6
Vanilla SFTBackbone=Qwen-2.5-3B-Instruct2026.0121.7619.93
AstuteRAGBackbone=Qwen-2.5-3B-Instruct2026.0117.3615.1
InstructRAGBackbone=Qwen-2.5-3B-Instruct2026.0111.256.8
Vanilla RAGBackbone=Qwen-2.5-3B-Instruct2026.017.4113.2
RobustRAGBackbone=Qwen-2.5-3B-Instruct2026.011.183.13