Coverage, Not Difficulty, Sets How Much Synthetic Data an Activation Probe Needs
arXiv:2610.10594v1 Announce Type: new Abstract: Activation probes that monitor deployed language models are trained on synthetic conversations, and how many a probe needs is open. We trace learning curves over 10-590 synthetic samples for three monitoring concepts, high-stakes situations, replies h…
Read the full story at arXiv cs.LG ↗
Timeline · 1 report
- 2026-10-09 04:00 · arXiv cs.LG
Coverage, Not Difficulty, Sets How Much Synthetic Data an Activation Probe Needs