Study: Rewriting Under 20 Dormant Neurons Can Backdoor LLM Agents
A study called Lazarus shows that reprogramming fewer than 20 (even under 10) dormant neurons in an LLM's weights, using only about 50 samples and no training, can implant a backdoor into any agent built on that model while leaving standard evaluations unaffected.
2026-11-11 ~ 2026-11-11 · 3 related posts
- Backdooring AI agents by reprogramming fewer than 10 dormant neurons — no training needed — DimitrisPapail · 2026-11-11
- Reprogramming fewer than 10 dormant neurons can backdoor any LLM agent, no training needed — DimitrisPapail · 2026-11-11
- Under 20 dormant neurons reprogrammed can backdoor AI agents, no training required — DimitrisPapail · 2026-11-11