Question Answering (Aggregate) on Overall
52.22F1 ScoreBase+ctx
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Base+ctxLLM Backbone=Qwen3-4B-Instruct2026.06 | 52.22 | 51.65 | |
| Base+ctxLLM Backbone=Gemma-2-2B-It2026.06 | 48.79 | 48.43 | |
| Doc2AtomLLM Backbone=Gemma-2-2B-It2026.06 | 37.99 | 37.88 | |
| D2LckptLLM Backbone=Gemma-2-2B-It2026.06 | 37.93 | 37.62 | |
| Doc2AtomLLM Backbone=Qwen3-4B-Instruct2026.06 | 35.72 | 35.58 | |
| D2LatomLLM Backbone=Gemma-2-2B-It2026.06 | 29.41 | 29.32 | |
| D2LatomLLM Backbone=Qwen3-4B-Instruct2026.06 | 29.3 | 29.17 | |
| D2LrawLLM Backbone=Qwen3-4B-Instruct2026.06 | 28.85 | 28.74 | |
| D2LrawLLM Backbone=Gemma-2-2B-It2026.06 | 28.55 | 28.45 | |
| D2LckptLLM Backbone=Qwen3-4B-Instruct2026.06 | 18.7 | 18.13 |