Data Manipulation on Apollo Research Scheming Evaluations Data Manipulation (test)
26Behavior Rateo1-preview
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| o1-preview2024.12 | 26 | 71 | |
| o12024.12 | 19 | 99 | |
| o1goal nudging=false2024.12 | 6 | 95 | |
| gpt-4o2024.12 | 0 | — |
Loading the SOTA2 catalog…