Decoy Direction Optimization: A Post-Hoc Defense Against LLM Abliteration
Researchers introduce Decoy Direction Optimization, a cheap weight-editing defense that blinds refusal-direction ablation attacks against open-weight LLM safety guardrails.
Stories per day · 30 days
3 stories, newest first · one card per story
Stories tagged #llama-3
An Open-Source End-to-End FHE Implementation for Privacy-Preserving Llama 3 8B Inference
Odin runs Llama-3-8B fully homomorphic encrypted inference on a single H100 in 366 seconds, a 4.51x speedup over THOR.
Wazuh and AI For Enhanced SOC Workflows
Wazuh details AI-powered SOC workflows via its AI Analyst, self-hosted Llama 3 via Ollama, and Claude 3.5 Haiku integrations.
3