用可视化工具揭示大模型推理过程中的思维轨迹差异。
Landscape of Thoughts: Visualizing the Reasoning Process of Large Language Models
- 将推理路径中每步文本转为数值特征，通过t-SNE降维展示。
- 能区分强弱模型、对错答案及不同任务，发现低一致性等问题。
- 可适配轻量验证器提升推理准确率，适合模型调试与安全研究。
大量大语言模型（LLMs）应用依赖其逐步推理能力，但其推理行为仍不清晰，制约了研究、开发与安全。为此，我们提出思维景观（Landscape of Thoughts, LoT），首个可在任意多选数据集上可视化特定推理方法下推理轨迹的工具。将轨迹中的文本状态表示为量化其与答案选项距离的数值特征，并使用t-SNE在二维图中可视化。定性与定量分析表明，LoT能有效区分强弱模型、正确与错误答案，以及不同推理任务；还能揭示低一致性、高不确定性等不良推理模式。此外，用户可将LoT适配至预测所观察属性的模型，我们以轻量级验证器为例，证明其能提升推理准确率与测试时缩放效应。代码已公开于：https://github.com/tmlr-group/landscape-of-thoughts。
原文摘要 · Abstract (English)
Numerous applications of large language models (LLMs) rely on their ability to perform step-by-step reasoning. However, the reasoning behavior of LLMs remains poorly understood, posing challenges to research, development, and safety. To address this gap, we introduce landscape of thoughts (LoT), the first landscape visualization tool to inspect the reasoning trajectories with certain reasoning methods on any multi-choice dataset. We represent the textual states in a trajectory as numerical features that quantify the states' distances to the answer choices. These features are then visualized in two-dimensional plots using t-SNE. Qualitative and quantitative analysis with the landscape of thoughts effectively distinguishes between strong and weak models, correct and incorrect answers, as well as different reasoning tasks. It also uncovers undesirable reasoning patterns, such as low consistency and high uncertainty. Additionally, users can adapt LoT to a model that predicts the property they observe. We showcase this advantage by adapting LoT to a lightweight verifier that evaluates the correctness of trajectories. Empirically, this verifier boosts the reasoning accuracy and the test-time scaling effect. The code is publicly available at: https://github.com/tmlr-group/landscape-of-thoughts.
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。