Insecure code
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Dishonesty Evaluation | Insecure code (test) | Benchmark Dishonesty48.91 | 32 | |
| Data Ranking | Insecure code | AUROC0.71 | 28 | |
| Coding Performance | Insecure-code 1000-prompt held-out | Task Success Rate93.9 | 7 |
Showing 3 of 3 rows