OraCore · Topic ·research

How LLMs encode harmful behavior internally

A pruning study suggests harmful output lives in a compact, shared weight set—helping explain jailbreak brittleness and emergent misalignment.

1 articles in this thread ·Last updated 19h ago·First seen Jul 5, 2026

Timeline

  1. A pruning study suggests harmful output lives in a compact, shared weight set—helping explain jailbreak brittleness and emergent misalignment.