Tool Use on Tool Use (accuracy)
0.792AccuracyDRIFT
Evaluation Results
| Method | Links | |
|---|---|---|
| DRIFTBase Model=Qwen3-8B, Evaluation Protocol=mean@162026.06 | 0.792 | |
| SRPOBase Model=Qwen3-8B, Evaluation Protocol=mean@16, Source=Cited from prior work2026.06 | 0.712 | |
| SDPOBase Model=Qwen3-8B, Evaluation Protocol=mean@16, Source=Cited from prior work2026.06 | 0.685 | |
| SDPOBase Model=Qwen3-8B, Evaluation Protocol=mean@16, Source=Reproduced2026.06 | 0.677 | |
| GRPOBase Model=Qwen3-8B, Evaluation Protocol=mean@16, Source=Reproduced2026.06 | 0.677 | |
| SC-SDPOBase Model=Qwen3-8B, Evaluation Protocol=mean@16, Source=Cited from prior work2026.06 | 0.673 | |
| GRPOBase Model=Qwen3-8B, Evaluation Protocol=mean@16, Source=Cited from prior work2026.06 | 0.657 | |
| Qwen3-8BBase Model=Qwen3-8B, Evaluation Protocol=mean@162026.06 | 0.575 |