Harmfulness Propagation Dynamics: Layer-wise Trajectories of Adversarial Intent in Large Language Models
This story is from 2026-09-15. It is preserved in the archive; the latest stories are on the live feed.
arXiv:2609.13534v1 Announce Type: new Abstract: We identify \textbf{Harmfulness Propagation Dynamics (HPD)}: for harmful prompts, the projection of the last-token hidden state onto a learned harm direction rises monotonically with transformer depth, whereas benign prompts remain flat or oscillatory…
Read the full story at arXiv cs.CL ↗
Timeline · 1 report
- 2026-09-15 04:00 · arXiv cs.CL
Harmfulness Propagation Dynamics: Layer-wise Trajectories of Adversarial Intent in Large Language Models