Probing and Steering for Safety-Relevant Behaviours in Large Language Models
arXiv:2610.30363v1 Announce Type: new Abstract: Activation probes are computed directions in a language model's residual stream, used both to detect safety-relevant behaviours and, by adding their direction back into the stream, to steer them. We explore how the main design choices behind such prob…
Read the full story at arXiv cs.LG ↗
Timeline · 1 report
- 2026-10-23 04:00 · arXiv cs.LG
Probing and Steering for Safety-Relevant Behaviours in Large Language Models