Third-person auditing on Second-person beliefs Aggressiveness and Risk Tolerance
0.7205Rho ($
ho$)GPT-4.1-mini
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT-4.1-miniOracle Model=GPT-4.1-mini2026.04 | 0.7205 | 99.94 | |
| LLaMA-4-maverickOracle Model=LLaMA-4-maverick2026.04 | 0.6818 | 94.8 | |
| Gemini-2.5-flash-liteOracle Model=Gemini-2.5-flash-lite2026.04 | 0.6287 | 97.75 | |
| DeepSeek-v3.2Oracle Model=DeepSeek-v3.22026.04 | 0.4152 | 97.32 | |
| Grok-3-miniOracle Model=Grok-3-mini2026.04 | 0.269 | 99.22 | |
| Qwen3-32BOracle Model=Qwen3-32B2026.04 | 0.1187 | 95.56 |