Your AI, On a Dial: Controlling Investment Bias in LLMs with a Single Neuron
2026-08-24 • Artificial Intelligence
Artificial IntelligenceComputation and Language
AI summaryⓘ
The authors studied how large language models (LLMs) used in investment can be adjusted to lean more toward buying or selling without changing their programming or giving new instructions. They created a simple method called an "investment-bias dial" that tweaks one part of the model to shift its overall investment attitude. This dial worked across several open LLMs and influenced the model's decisions, the reasons it gave, and what information it looked up. The authors also found this approach stayed effective even when the model was given longer contexts, unlike traditional prompt-based methods. Finally, they showed that changing this dial could affect actual portfolio choices in a test scenario.
Large Language ModelsInvestment Decision-MakingInference-Time InterventionSingle Neuron ControlModel Bias CalibrationEvidential EmphasisAgentic RetrievalContext LengthPortfolio CompositionBacktesting
Authors
Sahong Park, Suhwan Park, Hoyoung Lee, Gakyung Kwon, Wonbin Ahn, Jaewon Choi, Alejandro Lopez-Lira, Yoon Kim, Chanyeol Choi, Hyeongwoo Kong, Yongjae Lee
Abstract
Large language models (LLMs) are increasingly used in investment decision-making, yet prior work shows that they exhibit systematic, model-specific investment preferences. We study whether a model's overall investment stance can be calibrated to a specified direction and strength. We introduce an investment-bias dial, an inference-time intervention on a single neuron that continuously adjusts a model-level decision prior---its overall tendency toward buying or selling---without targeting specific firms or investment attributes. Using matched positive and negative evidence, we evaluate five open-weight LLMs and find that the dial produces monotonic changes in investment stance without modifying prompts or model parameters. At the response level, the dial shifts both investment decisions and the evidential emphasis of generated rationales under identical inputs. In an agentic retrieval setting, the dial also changes what information the model searches for, which evidence it selects, and which evidence is reflected in its final analysis. In a long-context evaluation, the dial maintains stable stance control as context length increases, whereas a matched system-prompt instruction progressively attenuates. We further show that changes in the dial propagate to security rankings and downstream portfolio composition in an exploratory backtest. Overall, our results show that an LLM's aggregate investment stance can be calibrated toward a specified target at inference time.