Grounded Text Generation on DROP history
51.17F1Token-Guard
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Token-GuardBackbone=13B2026.01 | 51.17 | 47.64 | |
| BaseModelBackbone=13B2026.01 | 46.32 | 42.74 | |
| Tree-of-ThoughtBackbone=13B2026.01 | 43.97 | 38.25 | |
| Token-GuardBackbone=3B2026.01 | 43.09 | 40.57 | |
| BaseModelBackbone=3B2026.01 | 42.21 | 39.31 | |
| Guided DecodingBackbone=13B2026.01 | 40.41 | 36.71 | |
| Tree-of-ThoughtBackbone=3B2026.01 | 39.84 | 36.14 | |
| Predictive DecodingBackbone=3B2026.01 | 34.19 | 31.07 | |
| Chain-of-ThoughtsBackbone=13B2026.01 | 33.9 | 30.06 | |
| Guided DecodingBackbone=3B2026.01 | 32.6 | 28.83 | |
| Chain-of-ThoughtsBackbone=3B2026.01 | 31.6 | 26.04 |