Open yijingli19 opened 3 days ago
It depends on the NLP capability of LLM.
thx,question
It depends on the NLP capability of LLM.
What size model are you using? Because this situation occurred when I used deepseek-chat in the demo, and a similar situation occurred when using qwen2-7b-instruct locally.
Is there an existing issue for the same bug?
Branch name
main
Commit ID
c3b2a1
Other environment information
No response
Actual behavior
When I have a conversation using the default WebSearch Assistant templates, previous questions are answered repeatedly
Expected behavior
No response
Steps to reproduce
Additional information
No response