Medical Explanation Generation on UK Medicine Year 1 (full breakdown)
0.0671BLEUILearner-LLM
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| ILearner-LLMK=52026.05 | 0.0671 | — | 0.7863 | 0.8066 | 0.1219 | 3.2005 | |
| Qwen3-8B SFTStrategy=SFT, Backbone=Qwen3-8B2026.05 | 0.0458 | 0.8629 | 0.8466 | 0.7387 | 0.2457 | 2.47 | |
| Gemma4-E4B-it SFTStrategy=SFT, Backbone=Gemma4-E4B-it2026.05 | 0.0428 | 0.8622 | 0.8447 | 0.7556 | 0.2962 | 3.1411 | |
| RLearner-LLMBackbone=Gemma4-E4B2026.05 | 0.028 | 0.8487 | 0.8469 | 0.6531 | 0.391 | 3.1089 | |
| RLearner-LLMBackbone=Qwen32026.05 | 0.0223 | 0.8161 | 0.7946 | 0.8195 | 0.2104 | 2.96 | |
| RLearner-LLMBackbone=LLaMA-22026.05 | 0.0222 | 0.8308 | 0.8467 | 0.7766 | 0.4251 | 3.0147 | |
| SFTBackbone=LLaMA-2-13B2026.05 | 0.0136 | 0.8065 | 0.7799 | 0.7556 | 0.086 | 3.2561 |