Search papers, labs, and topics across Lattice.
TeamLLM is introduced, a multi-LLM framework that mimics human team role division (specializing in roles like Project Manager, Researcher, and Writer) to improve performance on complex, multi-step tasks. The authors create a new benchmark, CGPST, to evaluate LLMs on contextually-grounded and procedurally-structured tasks, revealing limitations of single-perspective LLMs. Experiments on CGPST demonstrate that TeamLLM significantly outperforms individual LLMs by leveraging its team-oriented collaboration strategy.
LLMs working as specialized teams dramatically outperform solo LLMs on complex, multi-step reasoning tasks, suggesting a new paradigm for LLM collaboration.
Recently, multi-Large Language Model (LLM) frameworks have been proposed to solve contextualized tasks. However, these frameworks do not explicitly emulate human team role division, which may lead to a single perspective, thereby weakening performance on multi-step contextualized tasks. To address this issue, we propose TeamLLM, a human-like Team-Oriented Multi-LLM Collaboration Framework. TeamLLM adopts four team roles with distinct division and employs a three-phase multi-LLM collaboration for multi-step contextualized tasks. To evaluate the effectiveness of TeamLLM on multi-step contextualized tasks, we propose Contextually-Grounded and Procedurally-Structured tasks (CGPST) and construct the CGPST benchmark. This benchmark has four core features: contextual grounding, procedural structure, process-oriented evaluation and multi-dimensional assessment. We evaluate ten popular LLMs on CGPST at overall-level, step-level, and dimension-level. Results show that TeamLLM substantially improves performance on CGPST. We release the benchmark with scenarios, full-process responses and human scores from ten LLMs. The code and data are available at https://anonymous.4open.science/r/TeamLLM-anonymous-C50E/.