AI RESEARCH
Conv-to-Bench: Evaluating Language Models Via User-Assistant Dialogues In Code Tasks
arXiv CS.CL
•
ArXi:2605.26440v1 Announce Type: new The rapid advancement of Large Language Models (LLMs) has outpaced the scalability of traditional evaluation benchmarks, which remain heavily dependent on labor-intensive expert curation. We address this bottleneck with Con-to-Bench, a multi-stage framework that automatically transforms authentic multi-turn user-assistant dialogues into structured, verifiable requirement checklists.