Fetching the paper…

Debate Only When Necessary: Adaptive Multiagent Collaboration for Efficient LLM Reasoning · Around