Variable Tracking on RULER 16k
70.56F1 ScoreRephrasing
Evaluation Results
| Method | Links | |
|---|---|---|
| RephrasingModel (Max Context)=Llama-3-8B (262k)2025.06 | 70.56 | |
| PCDModel (Max Context)=Llama-3-8B (1048k)2025.06 | 69.11 | |
| PCDModel (Max Context)=Llama-3-8B (262k)2025.06 | 69.04 | |
| DoLa-HighModel (Max Context)=Llama-3-8B (262k)2025.06 | 67.29 | |
| DoLa-LowModel (Max Context)=Llama-3-8B (262k)2025.06 | 66.48 | |
| BaseModel (Max Context)=Llama-3-8B (1048k)2025.06 | 65.21 | |
| MsPoEModel (Max Context)=Llama-3-8B (262k)2025.06 | 65.1 | |
| Beam-SearchModel (Max Context)=Llama-3-8B (262k)2025.06 | 65.08 | |
| BaseModel (Max Context)=Llama-3-8B (262k)2025.06 | 64.4 | |
| SegRModel (Max Context)=Llama-3-8B (262k)2025.06 | 0 |