Rule Explanation Generation on Annotated Rule-Explanation Data Phase 3
4.8CorrectnessGemini 2.0 Flash
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Gemini 2.0 FlashRule Structure=2 atoms2025.08 | 4.8 | 4.76 | 29.98 | |
| Gemini 2.0 FlashRule Structure=binary2025.08 | 4.7 | 4.71 | 27.77 | |
| Gemini 2.0 FlashRule Structure=mediator2025.08 | 4.69 | 4.63 | 26.92 | |
| Gemini 2.0 FlashRule Structure=all2025.08 | 4.67 | 4.7 | 27.19 | |
| Gemini 2.0 FlashRule Structure=concatenated2025.08 | 4.63 | 4.75 | 26.19 | |
| Gemini 2.0 FlashRule Structure=3 atoms2025.08 | 4.61 | 4.68 | 25.62 | |
| GPT-4o miniRule Structure=2 atoms2025.08 | 4.52 | 4.62 | 40.96 | |
| GPT-4o miniRule Structure=binary2025.08 | 4.5 | 4.58 | 33.52 | |
| GPT-4o miniRule Structure=concatenated2025.08 | 4.5 | 4.51 | 30.38 | |
| GPT-4o miniRule Structure=all2025.08 | 4.45 | 4.53 | 31.57 | |
| GPT-4o miniRule Structure=3 atoms2025.08 | 4.42 | 4.51 | 26.26 | |
| GPT-3.5 TurboRule Structure=binary2025.08 | 4.4 | 4.42 | 34.58 | |
| GPT-3.5 TurboRule Structure=2 atoms2025.08 | 4.38 | 4.43 | 34.08 | |
| GPT-3.5 TurboRule Structure=all2025.08 | 4.28 | 4.26 | 32.4 | |
| GPT-4o miniRule Structure=mediator2025.08 | 4.24 | 4.49 | 26.82 | |
| GPT-3.5 TurboRule Structure=3 atoms2025.08 | 4.22 | 4.17 | 31.46 | |
| GPT-3.5 TurboRule Structure=mediator2025.08 | 4.13 | 4.07 | 26.26 | |
| GPT-3.5 TurboRule Structure=concatenated2025.08 | 4.1 | 4.07 | 31.57 |