Sparks of Rationality: Do Reasoning LLMs Align with Human Judgment and Choice?
🤖 AI Summary
This study investigates whether large language models replicate human patterns of rationality and emotional bias in high-stakes decision-making. Through behavioral economics tasks and social norm experiments, the authors systematically evaluate model adherence to rational choice axioms and sensitivity to emotional interventions, employing chain-of-thought reasoning, in-context prompting (ICP), and representation-level steering (RLS). Findings reveal that explicit “thinking” significantly enhances model rationality, aligning choices closer to expected-value maximization. While RLS elicits more human-like affective responses, ICP produces stronger but less controllable effects. The results uncover an inherent trade-off between reasoning capability and emotional sensitivity, offering new insights for designing models that are better aligned with human values while maintaining controllability.
Application Category
📝 Abstract
Large Language Models (LLMs) are increasingly positioned as decision engines for hiring, healthcare, and economic judgment, yet real-world human judgment reflects a balance between rational deliberation and emotion-driven bias. If LLMs are to participate in high-stakes decisions or serve as models of human behavior, it is critical to assess whether they exhibit analogous patterns of (ir)rationalities and biases. To this end, we evaluate multiple LLM families on (i) benchmarks testing core axioms of rational choice and (ii) classic decision domains from behavioral economics and social norms where emotions are known to shape judgment and choice. Across settings, we show that deliberate"thinking"reliably improves rationality and pushes models toward expected-value maximization. To probe human-like affective distortions and their interaction with reasoning, we use two emotion-steering methods: in-context priming (ICP) and representation-level steering (RLS). ICP induces strong directional shifts that are often extreme and difficult to calibrate, whereas RLS produces more psychologically plausible patterns but with lower reliability. Our results suggest that the same mechanisms that improve rationality also amplify sensitivity to affective interventions, and that different steering methods trade off controllability against human-aligned behavior. Overall, this points to a tension between reasoning and affective steering, with implications for both human simulation and the safe deployment of LLM-based decision systems.
Problem
Research questions and friction points this paper is trying to address.
Large Language Models
rationality
human judgment
decision-making
behavioral biases
Innovation
Methods, ideas, or system contributions that make the work stand out.
rationality
affective steering
large language models
behavioral economics
representation-level steering
🔎 Similar Papers
No similar papers found.
