Program generation on expert-curated Biology dataset
70.26Program AccuracyTaNOS
Evaluation Results
| Method | Links | |
|---|---|---|
| TaNOSSupervision budget=100% Data2026.04 | 70.26 | |
| TaNOS w/o SSLSupervision budget=100% Data2026.04 | 68.85 | |
| TaNOSSupervision budget=10% Data2026.04 | 63.85 | |
| SFTSupervision budget=100% Data2026.04 | 62.3 | |
| TaNOS w/o SSLSupervision budget=10% Data2026.04 | 62.18 | |
| SFT + SSLSupervision budget=100% Data2026.04 | 61.48 | |
| TaNOS w/o SketchSupervision budget=100% Data2026.04 | 59.84 | |
| TaNOS w/o SketchSupervision budget=10% Data2026.04 | 58.2 | |
| SFTSupervision budget=10% Data2026.04 | 57.38 | |
| SFT + SSLSupervision budget=10% Data2026.04 | 56.56 |