BARREL：面向事实性与可靠性的边界感知推理大语言模型

BARREL: Boundary-Aware Reasoning for Factual and Reliable LRMs

May 18, 2025

作者: Junxiao Yang, Jinzhe Tu, Haoran Liu, Xiaoce Wang, Chujie Zheng, Zhexin Zhang, Shiyao Cui, Caishun Chen, Tiantian He, Hongning Wang, Yew-Soon Ong, Minlie Huang

cs.AI

摘要

近期，大型推理模型（LRMs）在数学与逻辑推理方面展现出了令人瞩目的能力。然而，当前的LRMs极少承认无知或回应“我不知道”，反而常常在表现出过度自信的同时给出错误答案，这引发了对其事实可靠性的担忧。在本研究中，我们识别出两种由过度思考导致的病态推理模式，它们助长了这种过度自信与错误答案：最后一刻的猜测和反复纠结的螺旋思维。为解决这些问题，我们提出了BARREL——一个旨在促进简洁且边界意识强的事实推理的新框架。实验表明，经过BARREL训练后，DeepSeek-R1-Distill-Llama-8B的可靠性从39.33%提升至61.48%，同时保持了与基于R1生成推理数据进行微调的模型相当的准确度。这些结果证明，我们的初步研究对于构建更加可靠且基于事实的系统二型LRMs具有启发意义。

English

Recent advances in Large Reasoning Models (LRMs) have shown impressive capabilities in mathematical and logical reasoning. However, current LRMs rarely admit ignorance or respond with "I don't know". Instead, they often produce incorrect answers while showing undue confidence, raising concerns about their factual reliability. In this work, we identify two pathological reasoning patterns characterized by overthinking that contribute to the overconfident and incorrect answers: last-minute guessing and second-thought spiraling. To address these issues, we propose BARREL-a novel framework that promotes concise and boundary-aware factual reasoning. Our experiments show that BARREL-training increases the reliability of DeepSeek-R1-Distill-Llama-8B from 39.33% to 61.48%, while still achieving accuracy comparable to models finetuned on reasoning data generated by R1. These results demonstrate that our pilot study is inspiring to build more reliable and factual System 2 LRMs.