Reasoning Court: Combining Reasoning, Action, and Judgment for Multi-Hop Reasoning
This addresses the issue of error propagation in complex reasoning tasks for users of LLMs, representing an incremental improvement over existing hybrid methods like ReAct.
The paper tackles the problem of hallucinations and reasoning errors in large language models during multi-hop tasks by introducing Reasoning Court, a framework that extends iterative reasoning-and-retrieval methods with a dedicated LLM judge to evaluate and select or synthesize answers, resulting in consistent outperformance of state-of-the-art few-shot prompting methods on benchmarks like HotpotQA, MuSiQue, and FEVER without task-specific fine-tuning.
While large language models (LLMs) have demonstrated strong capabilities in tasks like question answering and fact verification, they continue to suffer from hallucinations and reasoning errors, especially in multi-hop tasks that require integration of multiple information sources. Current methods address these issues through retrieval-based techniques (grounding reasoning in external evidence), reasoning-based approaches (enhancing coherence via improved prompting), or hybrid strategies combining both elements. One prominent hybrid method, ReAct, has outperformed purely retrieval-based or reasoning-based approaches; however, it lacks internal verification of intermediate reasoning steps, allowing potential errors to propagate through complex reasoning tasks. In this paper, we introduce Reasoning Court (RC), a novel framework that extends iterative reasoning-and-retrieval methods, such as ReAct, with a dedicated LLM judge. Unlike ReAct, RC employs this judge to independently evaluate multiple candidate answers and their associated reasoning generated by separate LLM agents. The judge is asked to select the answer that it considers the most factually grounded and logically coherent based on the presented reasoning and evidence, or synthesizes a new answer using available evidence and its pre-trained knowledge if all candidates are inadequate, flawed, or invalid. Evaluations on multi-hop benchmarks (HotpotQA, MuSiQue) and fact-verification (FEVER) demonstrate that RC consistently outperforms state-of-the-art few-shot prompting methods without task-specific fine-tuning.