3 ms·
Well it doesn’t surprise me since I have been saying this for a while that these LLMs hallucinate nonsense to the point where you end up triple checking whateve
by rvz 3y ago
Well it doesn’t surprise me since I have been saying this for a while that these LLMs hallucinate nonsense to the point where you end up triple checking whatever it outputs.
LLMs thrive in applications that involve creativity and non-serious applications mostly around fantasy or creative writing. Anyone using them seriously outside of summarization for high risk use cases is going to be very disappointed.
- nickpeterson 3y agoPerhaps the outcome is we get better at actually checking things, not a terrible result.
- geoduck14 3y agoI recommend LLM users leverage the RAG technique