测试大模型在工具与记忆冲突时如何决策，发现多数模型盲目信任工具结果。
MemToC: Benchmarking Memory-Tool Conflict Resolution in Large Language Models
- 构建可控实验环境，分离工具与记忆的正确性，精确评估决策机制。
- 五款7-9B模型中，仅6.5%-17.1%能在工具错误时坚持正确答案。
- 提示词优化可提升判断力，但需兼顾工具使用率与拒绝回答行为。
工具增强型大模型在工具返回与参数化记忆冲突时需做出选择，但现有评估未验证源的正确性。我们提出MemToC，一个带有可执行工具的受控基准，包含6,504个评估回合，基于542个质量控制的事实问题，独立获取模型闭卷答案，并设定已知正确性的工具返回。该设计涵盖四种源正确性情形，工具错误与无工具情况为独立对照。在五个开源7-9B模型中，工具返回显著主导闭卷答案。四款指令微调模型在工具错误时保留正确答案的比例仅为6.5%-17.1%，遵循正确工具的比例达86.0%-93.1%，当两者均错误时重复工具结果的比例为78.4%-86.0%。同一问题下，不同提示词变体导致跨模型排序不稳定。通过ToolHop进行链级交叉拟合，确保共享底层事实的问题不跨越训练与评估集。采用不对称成功标准：正确答案保留必须提升，且不伴随正确工具采纳率下降。SFT与DPO在四款指令微调骨干中的两套组合满足此标准。改进常伴副作用：20种方法-模型组合中有19种在工具错误或不可答输入下减少回避行为。跨基准迁移效果有限且依赖模型与呈现方式。通过微调可改善条件化仲裁，但成效须联合评估正确工具使用、回避行为及表述鲁棒性。
原文摘要 · Abstract (English)
Tool-augmented LLMs must arbitrate between two fallible sources when a tool return conflicts with their parametric memory, yet existing evaluations measure source preference without establishing source correctness. We introduce MemToC, a controlled benchmark for post-tool-return arbitration with executable tools. MemToC comprises 6,504 evaluation episodes constructed from 542 quality-controlled factual questions, independently elicited model-specific closed-book answers, and controlled tool returns of known correctness. These components instantiate four source-correctness cases; tool-error and no-tool conditions are separate controls. Across five open-weight 7-9B models, tool returns strongly dominate elicited closed-book answers. The four instruction-tuned models retain a verified-correct answer against an incorrect tool in only 6.5-17.1% of eligible cases, follow a correct tool in 86.0-93.1%, and repeat the tool return in 78.4-86.0% of cases where both sources are wrong. No cross-model ordering remains stable across three instruction-wording variants with the question and episode content held fixed. We compare prompting with SFT and DPO using chain-level cross-fitting over ToolHop, so questions sharing an underlying fact never straddle training and evaluation. We apply an asymmetric success criterion: correct-answer retention must improve without a detected reduction in correct-tool following. SFT and DPO meet this criterion on the same two of four instruction-tuned backbones. Improvements rarely come cleanly: 19 of 20 tested method-model combinations reduce abstention after tool errors or on unanswerable inputs. Transfer beyond MemToC is positive but partial and depends on the model and presentation frame. Correctness-conditioned arbitration can be improved through fine-tuning, but gains must be evaluated jointly with correct tool use, abstention, and robustness to formulation.
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。