Hacker News Reader

Grandparent is correct. This is manually trained in, nothing fundamental changed since 2022. LLMs have no intrinsic way of knowing.

Source: I train LLMs and push their limits.

View on HN

You are both correct: the post-training stage of most new LLMs involves their identity being trained in, so that they "know who they are" without the system prompt. Without this step, most LLMs will respond with whatever identity most dominates their pre-training / post-training data, which is likely to be ChatGPT given its sheer prevalence.

There's some interesting anecdotal work on this with regards to self-recognition: https://josiekins.me/ai-comics

But this is not "system prompt"

Reply on news.ycombinator.com