18 ms·
In this context, is it that the training data had this problem? or that the training data had some semblance of the problem that is semantically linked to? o
by kumarvvr 3y ago
In this context,
is it that the training data had this problem?
or that the training data had some semblance of the problem that is semantically linked to?
or is it that the training data has no mention of this, and the model came up with a perfectly logical answer??
- jonplackett 3y agoI’ve been using GPT4 so much recently. Like all day long in chatGPT and the api making other services, and to all intents and purposes, it seems to be able to reason and think things through that other LLMs cannot. I have no idea why or how that is. But it feels that way to me using it all the time. 3.5 want like that. Bard / Gemini does not feel like that.