Story

Weird Generalization and Inductive Backdoors: New Ways to Corrupt LLMs

joegibbs Friday, December 12, 2025
2 1
arxiv.org
Visit article Read on Hacker News Comments 1