adversarial-perturbation
2 live · 3 on record · last incident 4 days ago
Live2
- liveRAG SNAPRAG: Cost-efficient knowledge extraction attack on RAG systems via adversarial perturbations
Research · since 10 Nov 2026
Researchers demonstrated SNAPRAG, an attack utilizing adversarial perturbations to extract private knowledge bases from retrieval-augmented generation deployments. The method reconstructs stored documents using significantly fewer queries and lower computational resources than existing extraction techniques.
Do:Deploy query rate limiting and monitor repeated semantic queries across external knowledge retrieval pipelines.
AI - liveHush-Up Adversarial Perturbation Attacks against Watchdogs in Distributed LLM Inference
Research · since 3 Nov 2026
Researchers introduced Hush-Up, a stealthy availability attack exploiting tolerance bounds in decentralized large language model inference verification. The paper demonstrates how malicious workers can execute availability attacks while avoiding detection mechanisms.
AI
Earlier1
- closedMirage: Targeted Jailbreaks Against Post-Generation Safety Checkers
Research · since 23 Oct 2026
AI