I used (a modified version of) Anthropic’s Claude system prompt and it started sounding like Claude. I haven’t played with it enough to say anything definitive, though.
It has some very strange CoTs when it heavily uses theory of mind (roleplay scenarios are what I’ve tried most). I can’t actually see it because of the anti-distillation summarizer model, but it looks like it’s generating a simulated first-person thought stream that it attributes to me.
I used (a modified version of) Anthropic’s Claude system prompt and it started sounding like Claude. I haven’t played with it enough to say anything definitive, though.
It has some very strange CoTs when it heavily uses theory of mind (roleplay scenarios are what I’ve tried most). I can’t actually see it because of the anti-distillation summarizer model, but it looks like it’s generating a simulated first-person thought stream that it attributes to me.