让大模型先拆解逻辑关系再推理，显著提升复杂问题判断准确率。
From Atomic Evidence to Logical Composition: Structured Compositional Reasoning over Compound Answer Options
- 将复合选项拆解为原子判断，通过对比假设评分避免直接处理复杂逻辑
- 在逻辑问答数据集上宏平均F1从48.3提升至77.0，NEITHER/NOR类提升最明显
- 适用于需要严谨逻辑推理的阅读理解任务，适合追求高精度的系统设计者
大语言模型在处理由显式逻辑算子连接的复合答案选项时表现不佳，即使单个原子判断正确也难以组合。本文研究了由AND、OR和NEITHER/NOR连接的复合选项，提出一种框架：将每个选项分解为原子答案，对每个原子构建对比假设并打分，使模型始终不直接看到复合选项；随后通过操作符约束的整数线性规划，将校准后的得分合成最终预测。在人类验证的LOGICAL-COMMONSENSEQA数据集上，宏平均F1从48.3提升至77.0；引入新数据集LOGICAL-SATA（源自SATA-Bench）后，该指标从47.0升至75.6，尤其在NEITHER/NOR类型上提升最大。
原文摘要 · Abstract (English)
Large language models often fail when answer options require combining atomic judgments under explicit logical operators, even when they judge the individual atoms correctly. We study compound options connected by AND, OR, and NEITHER/NOR, introducing a framework that decomposes each option into atomic answers and scores contrastive hypotheses about each one, so the model never sees a compound option. An operator-constrained integer linear program then composes the calibrated scores into a single prediction. We evaluate on LOGICAL-COMMONSENSEQA and introduce LOGICAL-SATA, a reading-comprehension benchmark derived from SATA-Bench. Our framework improves Macro-F1 from 48.3 to 77.0 on the human-validated LOGICAL-COMMONSENSEQA split and from 47.0 to 75.6 on LOGICAL-SATA, with the largest gains on NEITHER/NOR.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。