๐ŸŽ‰ MASCOT accepted to EMNLP 2026 Main as an oral presentation

Aug 20, 2026ยท
Yiyang "Diana" Wang
Yiyang "Diana" Wang
ยท 1 min read

MASCOT, our framework for multi-perspective socio-collaborative companions, was accepted to the EMNLP 2026 Main Conference as an oral presentation. MASCOT uses bi-level optimization to keep agents in character and make multi-agent dialogue more complementary and less sycophantic. Check out the paper, project page, and code.