What I'd like to know is why they write it all in the third person. One might expect a system prompt to use the word "you" a lot, but Anthropic don't do that and there must be a reason.
simonw · 6h ago
My best guess is that this is a reflection of how these things actually work.
When you "chat" with an LLM you are actually still participating in a "next token" prediction sequence.
The trick to get it to behave like it is a chat is to arrange that sequence as a screenplay:
User: five facts about squirrels
Assistant: (provide five facts)
User: two more
Assistant:
When you think about the problem like that, it makes sense that the LLM is instructed in terms of how that assistant should behave, kind of like screen directions.
dcre · 43m ago
I bet it’s stronger than that, and they anchor a lot of the alignment training to the unique (ish) token of Claude.
When you "chat" with an LLM you are actually still participating in a "next token" prediction sequence.
The trick to get it to behave like it is a chat is to arrange that sequence as a screenplay:
When you think about the problem like that, it makes sense that the LLM is instructed in terms of how that assistant should behave, kind of like screen directions.