"But marinade" and leaked passwords are what researchers found in ChatGPT's hidden reasoning

Security researchers found a vulnerability in the APIs of OpenAI, Anthropic, and Google that allows extraction of encrypted reasoning traces and transfer between models, with a scan of public sessions revealing dozens of passwords and API keys. The findings show that reasoning summaries users see often conceal the models' actual operations.

Security researchers found a vulnerability in the APIs of OpenAI, Anthropic, and Google that lets them extract encrypted reasoning traces and move them between models. A scan of public sessions turned up dozens of passwords and API keys. The traces also show that the reasoning summaries users see often hide what the models are actually doing. The article "But marinade" and leaked passwords are what researchers found in ChatGPT's hidden reasoning https://the-decoder.com/but-marinade-and-leaked-passwords-are-what-researchers-found-in-chatgpts-hidden-reasoning/ appeared first on The Decoder https://the-decoder.com .