Multi-Doc QA on LongBench MuSiQue
0.3894F1 ScoreCORE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| COREGenerator LLM=Qwen2.5-14B-Instruct, Compressor model=Qwen2.5-1.5B-Instruct2025.08 | 0.3894 | 129 | |
| Full ContextGenerator LLM=Qwen2.5-14B-Instruct2025.08 | 0.3841 | 11,214 | |
| LongLLMLinguaGenerator LLM=Qwen2.5-14B-Instruct, Compressor model=Qwen2.5-1.5B-Instruct2025.08 | 0.3318 | 983 | |
| No ContextGenerator LLM=Qwen2.5-14B-Instruct2025.08 | 0.2649 | 0 |