Instruction Following on IF XLow Tier
64.4R RateGated GSPO
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Gated GSPORL Training=Completed, Gating=Yes2026.06 | 64.4 | 0 | |
| Free GSPORL Training=Completed, Gating=None2026.06 | 62.9 | 50 | |
| BaselineRL Training=Step 0, Gating=None2026.06 | 52.4 | 20 |