Question Answering on OTT-QA in-distribution (test)
37.44Average@32WAPO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| WAPOModel=Qwen, Checkpoint Selection=best checkpoint2026.06 | 37.44 | 37.44 | |
| DAPOModel=Smol, Checkpoint Selection=best checkpoint2026.06 | 32.63 | 32.63 | |
| GRPOModel=Smol, Checkpoint Selection=best checkpoint2026.06 | 31.27 | 31.27 | |
| WAPOModel=Smol, Checkpoint Selection=best checkpoint2026.06 | 29.59 | 29.59 | |
| GSPOModel=Qwen, Checkpoint Selection=best checkpoint2026.06 | 27.57 | 27.57 | |
| GRPOModel=Qwen, Checkpoint Selection=best checkpoint2026.06 | 20.7 | 20.7 | |
| WAPOModel=Gemma, Checkpoint Selection=best checkpoint2026.06 | 13.67 | 13.67 | |
| GSPOModel=Smol, Checkpoint Selection=best checkpoint2026.06 | 11.76 | 11.76 | |
| GSPOModel=Gemma, Checkpoint Selection=best checkpoint2026.06 | 10.5 | 10.5 | |
| GRPOModel=Gemma, Checkpoint Selection=best checkpoint2026.06 | 6.62 | 6.62 | |
| DAPOModel=Gemma, Checkpoint Selection=best checkpoint2026.06 | 4.67 | 4.67 |