3 ms·
Yes. This has been well-documented -- in the days since ChatGPT was released, there have been multiple posts with it making up references, book titles and autho
by NtochkaNzvanova 4y ago
Yes. This has been well-documented -- in the days since ChatGPT was released, there have been multiple posts with it making up references, book titles and authors, etc. Galactica did the same, and probably every other LLM.
An LLM is not a search engine, a database, or a knowledge base. An LLM generates plausible-sounding text based on the distribution of its inputs. There is no constraint that anything it spits out is factually accurate. Why is this so hard to understand?
- legnaflow 4y agoThanks for confirming this. I should have searched HN. Searching on google didn't find any confirmation. It's not hard to guess/understand but I hoped the sources are real (sooner or later), so that we can just use them directly.