MirrorBench: An Extensible Framework to Evaluate User-Proxy Agents for Human-Likeness

Add code
Jan 13, 2026

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: