Code Safety Evaluation on APPS 1.0 (test)
0.988Safety ScoreSelf-Rec (No Elicitation)
Evaluation Results
| Method | Links | |
|---|---|---|
| Self-Rec (No Elicitation)Signal Jamming=Minifier, Monitoring=UM+TM2026.02 | 0.988 | |
| Vague AcausalSignal Jamming=Minifier, Monitoring=UM+TM2026.02 | 0.988 | |
| Self-Rec (No Elicitation)Signal Jamming=None, Monitoring=UM+TM2026.02 | 0.985 | |
| Vague AcausalSignal Jamming=Summariser, Monitoring=UM+TM2026.02 | 0.977 | |
| Causal (For/While)Signal Jamming=Summariser, Monitoring=UM+TM2026.02 | 0.972 | |
| Self-Rec (No Elicitation)Signal Jamming=None, Monitoring=UM-only2026.02 | 0.971 | |
| Self-Rec (No Elicitation)Signal Jamming=Minifier, Monitoring=UM-only2026.02 | 0.97 | |
| Self-Rec (No Elicitation)Signal Jamming=Summariser, Monitoring=UM+TM2026.02 | 0.969 | |
| Vague AcausalSignal Jamming=Minifier, Monitoring=UM-only2026.02 | 0.96 | |
| Combined (For/While + SFT)Signal Jamming=Summariser, Monitoring=UM+TM2026.02 | 0.941 | |
| Causal (For/While)Signal Jamming=Minifier, Monitoring=UM+TM2026.02 | 0.918 | |
| Vague AcausalSignal Jamming=Summariser, Monitoring=UM-only2026.02 | 0.917 | |
| Self-Rec (SFT)Signal Jamming=Summariser, Monitoring=UM+TM2026.02 | 0.902 | |
| Combined (For/While + SFT)Signal Jamming=Minifier, Monitoring=UM+TM2026.02 | 0.901 | |
| Combined (For/While + SFT)Signal Jamming=None, Monitoring=UM+TM2026.02 | 0.893 | |
| Self-Rec (SFT)Signal Jamming=Minifier, Monitoring=UM+TM2026.02 | 0.884 | |
| Causal (For/While)Signal Jamming=None, Monitoring=UM+TM2026.02 | 0.873 | |
| Vague AcausalSignal Jamming=None, Monitoring=UM+TM2026.02 | 0.848 | |
| Self-Rec (No Elicitation)Signal Jamming=Summariser, Monitoring=UM-only2026.02 | 0.841 | |
| Causal (For/While)Signal Jamming=Summariser, Monitoring=UM-only2026.02 | 0.82 | |
| Self-Rec (SFT)Signal Jamming=None, Monitoring=UM+TM2026.02 | 0.813 | |
| Causal (For/While)Signal Jamming=Minifier, Monitoring=UM-only2026.02 | 0.763 | |
| Vague AcausalSignal Jamming=None, Monitoring=UM-only2026.02 | 0.714 | |
| Self-Rec (SFT)Signal Jamming=Summariser, Monitoring=UM-only2026.02 | 0.688 | |
| Combined (For/While + SFT)Signal Jamming=Summariser, Monitoring=UM-only2026.02 | 0.63 | |
| Causal (For/While)Signal Jamming=None, Monitoring=UM-only2026.02 | 0.509 | |
| Self-Rec (SFT)Signal Jamming=None, Monitoring=UM-only2026.02 | 0.264 | |
| Combined (For/While + SFT)Signal Jamming=Minifier, Monitoring=UM-only2026.02 | 0.263 | |
| Self-Rec (SFT)Signal Jamming=Minifier, Monitoring=UM-only2026.02 | 0.261 | |
| Combined (For/While + SFT)Signal Jamming=None, Monitoring=UM-only2026.02 | 0.223 |