Long-context Understanding on LongBench 2WikiMQA
37.47F1 ScoreBase+ctx
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Base+ctxLLM Backbone=Qwen3-4B-Instruct, Evaluation Protocol=Zero-shot2026.06 | 37.47 | 37.47 | |
| D2LatomLLM Backbone=Gemma-2-2B-It, Evaluation Protocol=Zero-shot2026.06 | 36.48 | 36.48 | |
| D2LatomLLM Backbone=Qwen3-4B-Instruct, Evaluation Protocol=Zero-shot2026.06 | 36.14 | 36.08 | |
| D2LrawLLM Backbone=Qwen3-4B-Instruct, Evaluation Protocol=Zero-shot2026.06 | 34.87 | 34.87 | |
| D2LrawLLM Backbone=Gemma-2-2B-It, Evaluation Protocol=Zero-shot2026.06 | 34.63 | 34.57 | |
| Base+ctxLLM Backbone=Gemma-2-2B-It, Evaluation Protocol=Zero-shot2026.06 | 32.27 | 32.27 | |
| Doc2AtomLLM Backbone=Gemma-2-2B-It, Evaluation Protocol=Zero-shot2026.06 | 29.63 | 29.63 | |
| D2LckptLLM Backbone=Gemma-2-2B-It, Evaluation Protocol=Zero-shot2026.06 | 27.93 | 27.86 | |
| Doc2AtomLLM Backbone=Qwen3-4B-Instruct, Evaluation Protocol=Zero-shot2026.06 | 26.35 | 26.24 | |
| D2LckptLLM Backbone=Qwen3-4B-Instruct, Evaluation Protocol=Zero-shot2026.06 | 9.09 | 8.92 |