时间偏好概念及其在大语言模型中的功能
Temporal Preference Concepts and their Functions in a Large Language Model
AI总结：
通过因果定位和激活修补，本文发现大语言模型在中间到上层节点编码时间偏好几何结构，且行为分析表明模型对未来折扣比人类更平缓，但偏好不稳定，可通过引导向量调控。
摘要：
大语言模型（LLMs）越来越多地被部署用于需要在近期收益与长期后果之间权衡的决策，然而关于它们如何在内部表示或解决这些权衡，我们知之甚少。在这项工作中，我们通过因果定位了一个蒸馏LLM（Qwen3-4B-Instruct-2507）中时间偏好的底层子图，通过来自梯度归因和激活修补的汇聚证据识别了中上层节点。我们发现时间跨度的几何结构在预期局部层的残差流中被编码。行为分析表明，未干预的LLM对未来折扣的陡峭程度比人类低几倍，但这种偏好跨上下文不稳定，这促使我们进行显式控制而非隐式依赖训练。最后，我们发现有暗示性证据表明引导向量可以改变时间偏好。我们的工作展示了机械可解释性如何使我们更接近对LLM规划和推理方式的可靠控制。
英文摘要：
Large Language Models (LLMs) are increasingly being deployed to make decisions that require trading off near-term gains against long-term consequences, yet little is known about how they internally represent or resolve these tradeoffs. In this work, we causally localize an underlying subgraph for temporal preference in a distilled LLM (Qwen3-4B-Instruct-2507), identifying mid-to-upper-layer nodes through converging evidence from gradient-based attribution and activation patching. We find that the geometry of time horizon is encoded in the residual stream at the expected localized layers. A behavioral analysis reveals that unintervened LLMs discount the future several times less steeply than humans, yet this preference is unstable across contexts, motivating explicit control rather than implicit reliance on training. Finally, we find suggestive evidence that steering vectors can shift temporal preference. Our work demonstrates how mechanistic interpretability can bring us closer to reliable control over how LLMs plan and reason
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
