LLM Model-Swapping Trick Can Expose AI Reasoning Traces

2 points | by ashurandi 10 hours ago