Claude shows the authentic chain of thought (unless the system flags the COT as unsafe, in which case the user will be shown an encrypted version). It sounds from an announcement tweet like Gemini does as well, but I couldn’t find anything definitive in the docs for that one.
Claude shows the authentic chain of thought (unless the system flags the COT as unsafe, in which case the user will be shown an encrypted version). It sounds from an announcement tweet like Gemini does as well, but I couldn’t find anything definitive in the docs for that one.
Just wanted to follow up on this; I’ve read more Gemini COTs since then, and I currently have the strong impression that they’re summarized.