Replies: 1 comment 2 replies
-
|
It does in Native Function Calling Mode (model manage the calls to tools), most of the actual models manage its (smaller ones it it have it also is problematic) in Default Mode it doesn't (multitool calls), I made a PR (#17898 ) that do it, sequential tool calls in default mode. |
Beta Was this translation helpful? Give feedback.
2 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Check Existing Issues
Problem Description
Right now it can handle only basic tool calls that seem to be executed perhaps on side by some other LLM / tool and then just presents the results back to the model.
Desired Solution you'd like
I would like to see approach more similar to "thinking" process of ChatGPT - where the model is able to recursively run multiple tools (like web search), evaluate results, open more links as needed, and recursively run more tools, thinks and evaluates before presenting user with comprehensive and factually correct verified output.
Basically make it so that model can execute some tool, evaluate its outputs and decide if it wants to run another tool or if it's already ready to get back to the user, right now it's always at best single tool call.
Alternatives Considered
No response
Additional Context
No response
Beta Was this translation helpful? Give feedback.
All reactions