Entity · paper

Reason-Mediated Behavioral Models for Auditing LLM Social Simulators

paperactiveprovisionalreason-mediated-behavioral-models-for-auditing-llm-social-simulators-72be28c4·1 events·first seen 2d ago

Aliases: Reason-Mediated Behavioral Models for Auditing LLM Social Simulators

More like this (12)

Beyond Third-Person Audits: Situated Interaction Auditing for User-Centered LLM Bias Research Revising Context, Shifting Simulated Stance: Auditing LLM-Based Stance Simulation in Online Discussions Will Scaling Improve Social Simulation with LLMs?LLM Detection as an Intervention: Downstream Impact under Strategic User Behavior Moral Safety in LLMs: Exposing Performative Compliance with Puzzled Cues From Representations to Behaviors: Exploring the Person-Situation-Behavior Triad in LLMs Beyond Sycophancy: Structured Resistance and Compliance in LLM Moral Reasoning Situated Interaction Auditing Which Models Are Our Models Built On? Auditing Invisible Dependencies in Modern LLMs From Plausible to Actionable: A Position on LLM Self-Explanations What Do Safety-Aligned LLMs Learn From Mixed Compliance Demonstrations?Opaque Epistemic Mediation: How LLM Deployment Configurations Shape the Validation of Pseudo-Science

Recent events (1)

5arXiv · cs.AI·2d ago·source ↗

Reason-Mediated Behavioral Models for Auditing LLM Social Simulators

A new arXiv paper proposes an evaluation framework for LLMs used as social simulators, arguing that matching final outcomes is insufficient — simulators must also replicate the underlying rationale-derived reasoning paths. Using a 94-person sunscreen concept test, the authors map open-ended human rationales into signed 'reason states' and show that LLM-simulated reasons, while superficially plausible, frequently mirror the stimulus material rather than recovering the respondent's actual acceptance or rejection logic. The framework provides an interpretable audit for whether LLM social simulators are reasoning correctly, not just producing correct-looking outputs.

Evaluation and Benchmarking AI Safety Research Reason-Mediated Behavioral Models for Auditing LLM Social Simulators