Stealing Reasoning Traces from Proprietary LLM APIs

626 points

by quantumgarbage

21 hours ago

283 comments