用脑扫描技术解析大模型的经济预测逻辑，可调风险偏好与情绪倾向。
A Financial Brain Scan of the LLM
- 通过概念映射定位模型推理中的关键语义因子
- 实现风险厌恶/乐观等偏见的可控调节且不影响性能
- 适合社科研究者用于分析与矫正模型认知偏差
计算机科学的新方法使对大语言模型（LLMs）进行“脑扫描”成为可能，能够识别引导其推理的自然语言概念，并在保持其他因素不变的情况下对模型进行引导。我们展示该方法可将模型生成的经济预测映射到情感、技术分析和时机等概念，并计算其相对重要性，且不降低模型表现。同时，我们证明可通过调节使模型更具或更少风险规避性、乐观或悲观，从而帮助研究人员纠正或模拟认知偏见。该方法具有透明性、轻量级和可复现性，适用于社会科学的实证研究。
原文摘要 · Abstract (English)
Emerging techniques in computer science make it possible to "brain scan" large language models (LLMs), identify the plain-English concepts that guide their reasoning, and steer them while holding other factors constant. We show that this approach can map LLM-generated economic forecasts to concepts such as sentiment, technical analysis, and timing, and compute their relative importance without reducing performance. We also show that models can be steered to be more or less risk-averse, optimistic, or pessimistic, which allows researchers to correct or simulate biases. The method is transparent, lightweight, and replicable for empirical research in the social sciences.
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。