writinwaters

Results 37 comments of writinwaters

Hiya, here's the latest release notes for https://ragflow.io/docs/dev/release_notes#v0170. If you can't find the **Reasoning** toggle and the **Tavily API key** field, then it's not yet integrated into the demo.

@ld-xy Does #5114 work for you?

Do you mean "increase the number of the retrieved chunks"? If yes, lower the similarity threshold.

> > Do you mean "increase the number of the retrieved chunks"? If yes, lower the similarity threshold. > > I only find "the similarity threshold" in "retrieval test", is...

Your RAPTOR uses the system's default chat model. On the **Model providers** page, click **System Model Settings** to check if your system LLM is properly set.

If the right chunks can be retrieved, then the issue is not with RAG, but, more likely, with your chat assistant settings. 1: Have you changed the default value of...

Did you include {knowledge} in your prompt?