Long-context Question Answering on En.QA
36.75SubEMLogo-PO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Logo-POBackbone=Qwen2.5-14B-Instruct2025.08 | 36.75 | 32.12 | |
| SeaLong-POBackbone=Qwen2.5-14B-Instruct2025.08 | 36.47 | 25.93 | |
| LongMabBackbone=Qwen2.5-14B-Instruct2025.08 | 35.9 | 35.29 | |
| LongReward-POBackbone=Qwen2.5-14B-Instruct2025.08 | 35.61 | 27.1 | |
| Vanilla LLMBackbone=Qwen2.5-14B-Instruct2025.08 | 35.04 | 24.47 | |
| LongFaith-POBackbone=Qwen2.5-14B-Instruct2025.08 | 34.76 | 30.93 | |
| SeaLong-POBackbone=Qwen3-8B2025.08 | 32.86 | 33.54 | |
| LongAlpacaBackbone=Qwen2.5-14B-Instruct2025.08 | 32.76 | 26.76 | |
| LongMabBackbone=Llama-3.1-8B-Instruct2025.08 | 32.76 | 36.55 | |
| LongAlignBackbone=Qwen2.5-14B-Instruct2025.08 | 32.19 | 30.65 | |
| LongMab-SFTBackbone=Qwen2.5-14B-Instruct2025.08 | 31.91 | 28.42 | |
| LongMabBackbone=Qwen3-8B2025.08 | 31.43 | 33.84 | |
| LongMabBackbone=Qwen-2.5-7B-Instruct2025.08 | 30.48 | 31.88 | |
| Vanilla LLMBackbone=Qwen3-8B2025.08 | 30.29 | 31.35 | |
| Logo-POBackbone=Qwen3-8B2025.08 | 30.29 | 30.75 | |
| LongMab-SFTBackbone=Qwen3-8B2025.08 | 29.43 | 32.36 | |
| LongAlpacaBackbone=Qwen-2.5-7B-Instruct2025.08 | 29.34 | 23.82 | |
| LongReward-POBackbone=Qwen-2.5-7B-Instruct2025.08 | 28.2 | 21.77 | |
| SeaLong-POBackbone=Qwen-2.5-7B-Instruct2025.08 | 28.2 | 16.95 | |
| LongReward-POBackbone=Qwen3-8B2025.08 | 28 | 29.01 | |
| Logo-POBackbone=Qwen-2.5-7B-Instruct2025.08 | 28 | 28.47 | |
| LongAlpacaBackbone=Llama-3.1-8B-Instruct2025.08 | 27.35 | 30.58 | |
| SeaLong-POBackbone=Llama-3.1-8B-Instruct2025.08 | 27.06 | 32.06 | |
| Logo-POBackbone=Llama-3.1-8B-Instruct2025.08 | 26.67 | 32.69 | |
| Vanilla LLMBackbone=Llama-3.1-8B-Instruct2025.08 | 26.49 | 32.19 | |
| LongMab-SFTBackbone=Qwen-2.5-7B-Instruct2025.08 | 26.29 | 28.01 | |
| LongReward-POBackbone=Llama-3.1-8B-Instruct2025.08 | 26.21 | 24.2 | |
| Vanilla LLMBackbone=Qwen-2.5-7B-Instruct2025.08 | 25.64 | 22.38 | |
| LongAlpacaBackbone=Qwen3-8B2025.08 | 25.43 | 27.69 | |
| LongAlignBackbone=Qwen-2.5-7B-Instruct2025.08 | 25.07 | 25.61 | |
| LongAlignBackbone=Qwen3-8B2025.08 | 24.86 | 26.86 | |
| LongFaith-POBackbone=Llama-3.1-8B-Instruct2025.08 | 24.3 | 22.35 | |
| LongFaith-POBackbone=Qwen3-8B2025.08 | 24.29 | 28.08 | |
| LongAlignBackbone=Llama-3.1-8B-Instruct2025.08 | 23.93 | 24.85 | |
| LongMab-SFTBackbone=Llama-3.1-8B-Instruct2025.08 | 23.93 | 27.33 | |
| LongFaith-POBackbone=Qwen-2.5-7B-Instruct2025.08 | 23.07 | 18.55 |