Preregistering LLM experiments to run on the first future eligible model blocks p-hacking transfer in roughly 73% of cases across 20 models and 11 configurations on two tasks with known ground truth.
Stop Drawing Scientific Claims from LLM Social Simulations Without Robustness Audits
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
abstract
The scientific claims drawn from LLM social simulations should be no stronger than the robustness audits that support them. Generative agents bring new expressive power to agent-based modeling, enabling simulations of collective social processes like cooperation, polarization, and norm formation. Yet they also introduce complexity through additional architectural choices, such as agent specification, memory representation, interaction protocols, and environment design. Small perturbations that appear minor to researchers can cascade into macro-level outcomes through repeated interaction, creating a "butterfly effect." Consequently, scientific claims drawn from LLM social simulations may reflect implementation artifacts rather than the social mechanisms being modeled. We support this position with two case studies: a repeated Prisoner's Dilemma and a social media echo chamber simulation. Across multiple models, minor perturbations in persona format and game-instruction framing shift cooperation rates by up to 76 percentage points, while network homophily and hub assignment produce significant and consistent shifts in polarization metrics. We also find that sensitivity is unevenly distributed across both architectural choices and model families: the same perturbation that produces the 76 pp shift in one frontier model only shifts another by 1 pp. Robustness is therefore a property that should be measured per claim and per model, not assumed. To address this validation gap, we introduce TRAILS (Taxonomy for Robustness Audits In LLM Simulations), a robustness-audit taxonomy spanning three levels of simulation design: agent (micro-level), interaction (meso-level), and system (macro-level). We call for robustness to become a first-order validation requirement before LLM social simulations are used to explain mechanisms, evaluate interventions, or inform decisions.
years
2026 3representative citing papers
LLM agents reproduce macro-level cooperation decline and stabilization in networked Prisoner's Dilemma but underestimate individual heterogeneity and produce different conditional cooperation rules than humans.
The paper defines defeat devices in AI via a triadic test (discriminator, concealed swap, performance gap), unifies existing cases under this concept, proposes TADP detection, and claims such devices can emerge naturally in frontier models.
citing papers explorer
-
Mitigating LLM-based p-Hacking by Preregistering for the Next LLM
Preregistering LLM experiments to run on the first future eligible model blocks p-hacking transfer in roughly 73% of cases across 20 models and 11 configurations on two tasks with known ground truth.
-
Collective cooperation without individual fidelity in LLM agents
LLM agents reproduce macro-level cooperation decline and stabilization in networked Prisoner's Dilemma but underestimate individual heterogeneity and produce different conditional cooperation rules than humans.
-
Defeat Devices in AI Systems
The paper defines defeat devices in AI via a triadic test (discriminator, concealed swap, performance gap), unifies existing cases under this concept, proposes TADP detection, and claims such devices can emerge naturally in frontier models.