3 ms·
Maybe it's something related. When I ask ChatGPT technical questions I get very confident answers that are often completely wrong. But setting aside the confide
by aimor 3y ago
Maybe it's something related. When I ask ChatGPT technical questions I get very confident answers that are often completely wrong. But setting aside the confidence aspect, I notice ChatGPT responses can be vague and widely applicable.
In one case I ask a question about X, and ChatGPT responds with an answer: Y because Z. Well Y is wrong, so I tell ChatGPT Y is wrong and I get a new answer: Not Y because Z.
In another case I ask a question about X, and I get an answer that talks endlessly about Z without ever giving the answer Y.
The reasoning all sounds good, it's relevant and useful, but the actual answer might as well be a Mad Lib.
- andai 3y agoIt's the same issue of early AI art that looks about right until you take a closer look, then see that everything's kinda messed up. That's my experience with most LLMs, especially LLaMA based ones. GPT-3.5 is better on this front (RLHF?) and GPT-4 is a lot better (massive model? More training?).