Separated expert teams improve large language model question answering accuracy
SEPAL: Separated Expert Pairs with Answer-Level Fusion for Reliable LLM Collaboration
Computation and LanguageArtificial IntelligenceMachine Learning
Summary
Answering questions with large language models (LLMs) can be improved by letting multiple models work together and check each other's reasoning. The authors introduce SEPAL, a system where three smaller teams each focus on a different part of the process: reasoning, finding evidence, and verifying answers. By keeping feedback private within each team and only voting on the final answers, SEPAL avoids shared mistakes and improves accuracy. Their tests show better results across various LLMs and question sets.
What this means in practice
- •For ai product developers: Improve the accuracy of question answering features by integrating separated expert teams with private feedback and answer-level voting.$Commercial implications: Enables development of more reliable AI assistants and chatbots with higher-quality answers by reducing joint error propagation.
- •For nlp system engineers: Design large language model ensembles with specialized reasoning, grounding, and verification modules that collaborate without shared error feedback.
Authors
Weijie Ren, Yanwen Zhang, Hao Li, Zhuolin Qi, Hengyi Zhang, Naibo Wang
Abstract
Multi-agent collaboration lets large language models (LLMs) improve question answering through deliberation and feedback. Yet shared discussion couples correction with exposure to the same mistakes, which can erode the diversity needed for voting. Self-consistency offers sampling diversity without feedback, while single-pair Actor-Critic collaboration refines only one candidate. We introduce SEPAL, which assigns three private Actor-Critic teams to direct reasoning, evidence grounding, and verification. Role-specific training gives the teams different reasoning objectives beyond sampling variation. Each Critic guides revisions within its own team, preventing feedback from carrying errors across candidates. Once revision ends, majority voting combines only the final answers, keeping the reasoning histories separate until the decision. Across five open-weight backbones and five question-answering benchmarks, SEPAL improves mean accuracy by 1.81 percentage points over a matched single Actor-Critic pair, with improvements across all five backbones. Code is available at https://github.com/zhansan114514/SEPAL.