alt Hacker News"Prompts and completions are retained by the provider and are not used for training..."
I'm curious what the model provider is using the prompt/response pairs for, in that case. They aren't offering a model for free without their name on it for no reason.
Replies
I think this means its used for training, but there is some elaborate legal argument that allows them to claim its really not.
Like they dont pretrain on your chats but they do some transforms/RLHF/sentiment analysis. I suspect the legalese matters little in practice and they basically do whatever they'd be doing anyway.
Stealth Model, is this a CTF?
LLM needs to become more transparent, not less. Hence, this idea (and trend, possibly) is disgusting.
How can we even possibly verify 'Prompts and completions are retained by the provider and are not used for training...'? What if the training is done, but used internally?
Research, analytics, usage trends, etc. All still incredibly valuable for a company building and tuning an LLM; even if the data itself isn’t directly used in the training set.