[Feature]: Parse GLM5's mid-tool chain comments to appear as the 💬 inline comments instead of full messages #2636
Description
Problem or Use Case
I don't know if it's a quirk of the model or because im using nvidia nim to try out the model, but what would usually be displayed as little "💬 " blocks inline with tool comments seems to become full mid-action messages with GLM5, it would be nice if whatever GLM5 does differently could be parsed into the same little inline blocks as i've seen with other models since it's much neater and easier to parse in long running tool chains
This is with reasoning effort medium and shown, but hidden behaves the same
Example:
Proposed Solution
I don't know how much effort it would be and I don't really know how I could best capture raw output to help diagnose it but just adding an extra check for whatever GLM5 does differently so that it's mid-action responses were captured as the inline commentary blocks, or at least maybe as an additional option on top of the reasoning options if it isn't standard
Alternatives Considered
No response
Feature Type
Other
Scope
Small (single file, < 50 lines)
Contribution
- I'd like to implement this myself and submit a PR