I know this is not the theme of this post, but I wonder if there’s an LLM that doesn’t hallucinate when asked to summarize information of a group of documents. I tried Gpt4all for simple queries like finding out which documents mentioned a certain phrase. It often gave me filenames that didn’t actually exist. Hallucinating contents is one thing but making up data source is just horrible.
I know this is not the theme of this post, but I wonder if there’s an LLM that doesn’t hallucinate when asked to summarize information of a group of documents. I tried Gpt4all for simple queries like finding out which documents mentioned a certain phrase. It often gave me filenames that didn’t actually exist. Hallucinating contents is one thing but making up data source is just horrible.