Skip to content

Conversation

@reachtarunhere
Copy link
Collaborator

Please look at the notebook (open raw or download html) for charts displaying the results
The results folder contains the results for different thresholds.

Some interesting trends are observed that we can use fewer experts for domains like Humanities without losing much performance.

Factual questions almost never improve with more experts.

reachtarunhere and others added 25 commits November 12, 2025 22:25
Co-authored-by: aider (gemini/gemini-2.5-pro) <[email protected]>
Co-authored-by: aider (gemini/gemini-2.5-pro) <[email protected]>
Co-authored-by: aider (gemini/gemini-2.5-pro) <[email protected]>
Co-authored-by: aider (gemini/gemini-2.5-pro) <[email protected]>
Co-authored-by: aider (gemini/gemini-2.5-pro) <[email protected]>
Co-authored-by: aider (gemini/gemini-2.5-pro) <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants