build1 publisher
Eleven models encode 'no admissible answer' on an axis 85 degrees off safety refusal
A one-dimensional probe finds structural impossibility in the hidden state of instruction-tuned models from 1.7B to 70B parameters. The trained refusal pathway that guardrail work tunes reads an axis about 85 degrees away from it.
Publishers:arxiv.org
Reality
- Evidence58
- Adoption
- Insufficient
- Hype gap−12
- Incentives40
- Confidence52