OpenAI Compaction Flaw Let Researchers Bypass Hidden Reasoning Protections
// AI summary
Xint showed OpenAI's compaction feature could allow researchers to copy a model's hidden reasoning out in plain text, bypassing normal encrypted protections.
// from the source
OpenAI's compaction feature, which shrinks a long conversation to save space, could be turned into a way to copy a model's hidden reasoning out in plain text, security firm Xint has shown. OpenAI's reasoning models work through a problem before answering, and the service normally returns only the raw reasoning as an encrypted item the user cannot read. The method got past