Skip to content
Dispatch
Support
Send feedback
Revision history
Paper finds overall LLM evaluation accuracy can obscure prompt-specific errors in interactive settings
Original publish · no revisions.
← Back to article
Tweaks