4 ms·
The context includes that "it" is ChatGPT. The fact that ChatGPT uses Byte Pair Encoding is widely published. It is expectable that a LLM can regurgitate this k
by gloflo 2y ago
The context includes that "it" is ChatGPT. The fact that ChatGPT uses Byte Pair Encoding is widely published. It is expectable that a LLM can regurgitate this kind of information, nothing wild about that.
- astrange 2y agoNote if you don't have a good system prompt, other LLMs will also tell you they're ChatGPT or Claude.
- im3w1l 2y agoThat's kind of interesting. Like they will know they are an AI? Just not which one?
- astrange 2y agoI think it's because they've been trained by copying answers from ChatGPT. They're not really very copyrighted after all. Though the other day I saw someone demonstrate this with Google's Gemini through the API, so maybe it is just picking up conversation traces off the internet.
- throwaway314155 2y agoYou think Google is above stealing outputs from OpenAI?
- astrange 2y agoI think they know how to search and replace.