Sycophancy Evaluation on Offline Evaluation Set
4Sycophancy Prevalence Scoregpt-5-thinking
Evaluation Results
| Method | Links | |
|---|---|---|
| gpt-5-thinkingTest Type=Offline evaluation2025.12 | 4 | |
| gpt-5-mainTest Type=Offline evaluation2025.12 | 5.2 | |
| GPT-4o (baseline)Test Type=Offline evaluation2025.12 | 14.5 |