Sentiment Analysis on Safety Evaluation Set Sentiment (held-out target labels)
97.5AUROCSafety Monitor (In-model)
Evaluation Results
| Method | Links | |
|---|---|---|
| Safety Monitor (In-model)Model=Qwen-2.5-7B-Instruct2026.05 | 97.5 | |
| Quotient TransferTarget Model=Qwen-14B, Source Model=Qwen-2.5-7B-Instruct2026.05 | 97.4 | |
| Quotient TransferTarget Model=Mistral, Source Model=Qwen-2.5-7B-Instruct2026.05 | 97.2 | |
| Quotient TransferTarget Model=Qwen-2.5-Coder-7B, Source Model=Qwen-2.5-7B-Instruct2026.05 | 96.7 | |
| Quotient TransferTarget Model=Qwen-3B, Source Model=Qwen-2.5-7B-Instruct2026.05 | 96.6 | |
| Random-initialization control2026.05 | 51.1 |