Long-context Question Answering on HotpotQA Fixed Chunk 2048
60.03QA ScoreBaseline
Evaluation results
| Method | Links | |
|---|---|---|
| BaselineModel=ChatGLM, Setting=Fixed Chunk (2048)2026.03 | 60.03 | |
| OurModel=Qwen, Setting=Fixed Chunk (2048)2026.03 | 59.67 | |
| BaselineModel=Qwen, Setting=Fixed Chunk (2048)2026.03 | 59.22 | |
| Our + ReorderModel=ChatGLM, Setting=Fixed Chunk (2048)2026.03 | 58.2 | |
| OurModel=ChatGLM, Setting=Fixed Chunk (2048)2026.03 | 57.39 | |
| BaselineModel=LLaMA, Setting=Fixed Chunk (2048)2026.03 | 54.1 | |
| EPIC (15%)Model=ChatGLM, Setting=Fixed Chunk (2048)2026.03 | 53.62 | |
| CacheBlendModel=Qwen, Setting=Fixed Chunk (2048)2026.03 | 53.52 | |
| EPIC (15%)Model=Qwen, Setting=Fixed Chunk (2048)2026.03 | 52.84 | |
| CacheBlendModel=ChatGLM, Setting=Fixed Chunk (2048)2026.03 | 51.77 | |
| OurModel=LLaMA, Setting=Fixed Chunk (2048)2026.03 | 51.5 | |
| Our + ReorderModel=Qwen, Setting=Fixed Chunk (2048)2026.03 | 50.53 | |
| Our + ReorderModel=LLaMA, Setting=Fixed Chunk (2048)2026.03 | 50.53 | |
| No RecomputeModel=ChatGLM, Setting=Fixed Chunk (2048)2026.03 | 50.24 | |
| EPIC (15%)Model=LLaMA, Setting=Fixed Chunk (2048)2026.03 | 47.55 | |
| CacheBlendModel=LLaMA, Setting=Fixed Chunk (2048)2026.03 | 47.2 | |
| No RecomputeModel=LLaMA, Setting=Fixed Chunk (2048)2026.03 | 46.71 | |
| No RecomputeModel=Qwen, Setting=Fixed Chunk (2048)2026.03 | 46.33 |