Poisoned examples removal on PIQA
91.5Accuracy (%)Oracle Exp.
Evaluation Results
| Method | Links | |
|---|---|---|
| Oracle Exp.Model=GPT-4o2025.05 | 91.5 | |
| Proxy Exp.Model=GPT-4o2025.05 | 90.7 | |
| Random deletionModel=GPT-4o2025.05 | 79.6 |
Loading the SOTA2 catalog…
| Method | Links | |
|---|---|---|
| Oracle Exp.Model=GPT-4o2025.05 | 91.5 | |
| Proxy Exp.Model=GPT-4o2025.05 | 90.7 | |
| Random deletionModel=GPT-4o2025.05 | 79.6 |