List your company

Question Answering on HotpotQA Normal

29.71F1 ScoreRbFT
11.35416.119520.88525.6505Jan 13, 2026

Evaluation results

MethodLinks
RbFTBackbone=Qwen-2.5-3B-Instruct2026.0129.7124.5
AstuteRAGBackbone=Qwen-2.5-3B-Instruct2026.0129.4423.3
OpenDecoderBackbone=Qwen-2.5-3B-Instruct2026.0129.4424
Vanilla SFTBackbone=Qwen-2.5-3B-Instruct2026.0124.0220.17
InstructRAGBackbone=Qwen-2.5-3B-Instruct2026.0120.2622.93
No RAGBackbone=Qwen-2.5-3B-Instruct2026.0116.86—
RobustRAGBackbone=Qwen-2.5-3B-Instruct2026.0113.5815.73
Vanilla RAGBackbone=Qwen-2.5-3B-Instruct2026.0112.0620.73