distributed LLM inference

Mentions · 7d0
Mentions · 30d1
All mentions1
First seenOct 31, 2026
Last mentioned11d ago

Timeline

1 articles · newest first

Hush-Up attack halts distributed LLM inference

A single malicious worker forces premature LLM termination with small output perturbations that evade local activation-based verification.

arXiv cs.CROct 31, 16:38 UTC · 11d agoResearch

Appears with

LLMai model · 1

Entities are extracted by the model from each article. First seen Oct 31, 2026, last mentioned 11d ago. Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.