pith. sign in

•ASGUARDActivation Scaling:

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

fields

cs.AI 1

years

2025 1

verdicts

UNVERDICTED 1

representative citing papers

ASGuard: Activation-Scaling Guard to Mitigate Targeted Jailbreaking Attack

cs.AI · 2025-09-30 · unverdicted · novelty 6.0

ASGuard identifies tense-vulnerable attention heads via circuit analysis, trains precise scaling vectors on those activations, and applies them in preventative fine-tuning to reduce targeted jailbreaking success across four LLMs with minimal impact on utility or over-refusal.

citing papers explorer

Showing 1 of 1 citing paper.

  • ASGuard: Activation-Scaling Guard to Mitigate Targeted Jailbreaking Attack cs.AI · 2025-09-30 · unverdicted · none · ref 12

    ASGuard identifies tense-vulnerable attention heads via circuit analysis, trains precise scaling vectors on those activations, and applies them in preventative fine-tuning to reduce targeted jailbreaking success across four LLMs with minimal impact on utility or over-refusal.