Sentence-level length-controlled question answering on ELI5 (test)
0.08MAEICL+Feedback
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ICL+FeedbackBackbone=Qwen3-8B, Strategy=ICL+Feedback2026.01 | 0.08 | 92.8 | 7.69 | |
| FeedbackBackbone=LLaMA-3.1-8B-Instruct, Strategy=Feedback2026.01 | 0.11 | 95.75 | 7.05 | |
| FeedbackBackbone=Qwen3-8B, Strategy=Feedback2026.01 | 0.12 | 91.15 | 7.51 | |
| ICL+FeedbackBackbone=LLaMA-3.1-8B-Instruct, Strategy=ICL+Feedback2026.01 | 0.3 | 87.13 | 7.18 | |
| SFT+FeedbackBackbone=Qwen3-8B, Strategy=SFT+Feedback2026.01 | 0.4 | 78.6 | 8.11 | |
| SFT+FeedbackBackbone=LLaMA-3.1-8B-Instruct, Strategy=SFT+Feedback2026.01 | 1.54 | 44.4 | 7.45 | |
| Qwen3-8BBackbone=Qwen3-8B, Strategy=Baseline2026.01 | 1.82 | 48.67 | 7.57 | |
| SFTBackbone=Qwen3-8B, Strategy=SFT2026.01 | 2.21 | 19.18 | 8.18 | |
| ICLBackbone=Qwen3-8B, Strategy=ICL2026.01 | 2.74 | 34.57 | 7.71 | |
| SFTBackbone=LLaMA-3.1-8B-Instruct, Strategy=SFT2026.01 | 3.43 | 11.12 | 7.5 | |
| LLaMA-3.1-8B-InstructBackbone=LLaMA-3.1-8B-Instruct, Strategy=Baseline2026.01 | 3.52 | 33.45 | 7.26 | |
| ICLBackbone=LLaMA-3.1-8B-Instruct, Strategy=ICL2026.01 | 4.05 | 27.84 | 7.31 |