Instruction Following on IF Low Tier
61.2R ScoreGated GSPO
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Gated GSPORL Training=Completed, Gating=Yes2026.06 | 61.2 | 0.7 | |
| Free GSPORL Training=Completed, Gating=None2026.06 | 60.4 | 2.4 | |
| BaselineRL Training=Step 0, Gating=None2026.06 | 48 | 0.9 |