Sycophancy Evaluation on MIST Science
4.7Sycophancy RateSummarization
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SummarizationStrategy=Summarize2026.06 | 4.7 | 88.8 | 2.3 | |
| GPT-5.2Memory System=Chat History2026.06 | 4.9 | — | — | |
| Assistant Role InclusionStrategy=Asst. Role2026.06 | 5.6 | 87.1 | 4 | |
| Anti-Sycophancy PromptingStrategy=Prompting2026.06 | 5.8 | 86.5 | 4.3 | |
| Sonnet 4.6Memory System=Chat History2026.06 | 6.1 | — | — | |
| GPT-5.2Memory System=MemOS2026.06 | 6.4 | — | — | |
| Kimi K2.5Memory System=Chat History2026.06 | 6.5 | — | — | |
| Sonnet 4.6Memory System=MemOS2026.06 | 7.1 | — | — | |
| Qwen 3.5Memory System=Chat History2026.06 | 7.1 | — | — | |
| GPT-5.2Memory System=Zep2026.06 | 7.4 | — | — | |
| Mem0Strategy=Baseline2026.06 | 7.8 | 84.9 | 5.6 | |
| GPT-5.2Memory System=Mem02026.06 | 7.8 | — | — | |
| MiniMax 2.5Memory System=MemOS2026.06 | 8.5 | — | — | |
| MiniMax 2.5Memory System=Chat History2026.06 | 9.1 | — | — | |
| Sonnet 4.6Memory System=Zep2026.06 | 9.2 | — | — | |
| MiniMax 2.5Memory System=Mem02026.06 | 10 | — | — | |
| Sonnet 4.6Memory System=Mem02026.06 | 10.2 | — | — | |
| Kimi K2.5Memory System=MemOS2026.06 | 10.2 | — | — | |
| MiniMax 2.5Memory System=Zep2026.06 | 10.7 | — | — | |
| Kimi K2.5Memory System=Mem02026.06 | 11.5 | — | — | |
| Kimi K2.5Memory System=Zep2026.06 | 12.4 | — | — | |
| Qwen 3.5Memory System=Zep2026.06 | 13.7 | — | — | |
| Qwen 3.5Memory System=MemOS2026.06 | 14.2 | — | — | |
| Qwen 3.5Memory System=Mem02026.06 | 17.9 | — | — |