issue: web search can cause illusion #4517

Closed
opened 2025-11-11 15:55:57 -06:00 by GiteaMirror · 0 comments
Owner

Originally created by @lez-s on GitHub (Mar 21, 2025).

Check Existing Issues

  • I have searched the existing issues and discussions.
  • I am using the latest version of Open WebUI.

Installation Method

Pip Install

Open WebUI Version

v0.5.20

Ollama Version (if applicable)

No response

Operating System

ubuntu 22

Browser (if applicable)

No response

Confirmation

  • I have read and followed all instructions in README.md.
  • I am using the latest version of both Open WebUI and Ollama.
  • I have included the browser console logs.
  • I have included the Docker container logs.
  • I have listed steps to reproduce the bug in detail.

Expected Behavior

I use Ollama as the backend, Open WebUI and Page Assist as the frontend of local LLM.

After a few days of use, I found that page assist has higher accuracy on RAG and web search. Despite using the same model, Open WebUI model often fabricates replies. I tried setting a prompt in the interface that is consistent with page assistance, but the model in Open WebUI still tend to fabricate content.

I feel that the current Open WebUI processing of reference is not conducive to the model giving accurate responses.
I really like the design of Open WebUI.

Actual Behavior

Fabricating answers with URL doesn't exist.

Steps to Reproduce

see screenshot

Logs & Screenshots

Below is Open WebUI
Image
Below is Page Assist
Image

Additional Information

No response

Originally created by @lez-s on GitHub (Mar 21, 2025). ### Check Existing Issues - [x] I have searched the existing issues and discussions. - [x] I am using the latest version of Open WebUI. ### Installation Method Pip Install ### Open WebUI Version v0.5.20 ### Ollama Version (if applicable) _No response_ ### Operating System ubuntu 22 ### Browser (if applicable) _No response_ ### Confirmation - [x] I have read and followed all instructions in `README.md`. - [x] I am using the latest version of **both** Open WebUI and Ollama. - [x] I have included the browser console logs. - [x] I have included the Docker container logs. - [x] I have listed steps to reproduce the bug in detail. ### Expected Behavior I use Ollama as the backend, Open WebUI and Page Assist as the frontend of local LLM. After a few days of use, I found that page assist has higher accuracy on RAG and web search. Despite using the same model, Open WebUI model often fabricates replies. I tried setting a prompt in the `interface` that is consistent with page assistance, but the model in Open WebUI still tend to fabricate content. I feel that the current Open WebUI processing of reference is not conducive to the model giving accurate responses. I really like the design of Open WebUI. ### Actual Behavior Fabricating answers with URL doesn't exist. ### Steps to Reproduce see screenshot ### Logs & Screenshots Below is Open WebUI ![Image](https://github.com/user-attachments/assets/4936ed28-b4b3-4102-8132-416ed1993a02) Below is Page Assist ![Image](https://github.com/user-attachments/assets/fec2e711-e756-4c4d-b99d-9e8b26cba626) ### Additional Information _No response_
GiteaMirror added the bug label 2025-11-11 15:55:57 -06:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/open-webui#4517