CoMPosT: Characterizing and Evaluating Caricature in LLM Simulations
read the original abstract
Recent work has aimed to capture nuances of human behavior by using LLMs to simulate responses from particular demographics in settings like social science experiments and public opinion surveys. However, there are currently no established ways to discuss or evaluate the quality of such LLM simulations. Moreover, there is growing concern that these LLM simulations are flattened caricatures of the personas that they aim to simulate, failing to capture the multidimensionality of people and perpetuating stereotypes. To bridge these gaps, we present CoMPosT, a framework to characterize LLM simulations using four dimensions: Context, Model, Persona, and Topic. We use this framework to measure open-ended LLM simulations' susceptibility to caricature, defined via two criteria: individuation and exaggeration. We evaluate the level of caricature in scenarios from existing work on LLM simulations. We find that for GPT-4, simulations of certain demographics (political and marginalized groups) and topics (general, uncontroversial) are highly susceptible to caricature.
This paper has not been read by Pith yet.
Forward citations
Cited by 3 Pith papers
-
Marginal Alignment Does Not Guarantee Joint-Distribution Fidelity: An Official-Reference Audit of Nemotron-Personas-Korea with Cross-Locale Replication
An audit of one million Korean synthetic personas shows marginal demographic alignment does not preserve joint distributions, with three specific mismatches identified via a new Independence-Assumption Footprint method.
-
Disentangling Interaction and Bias Effects in Opinion Dynamics of Large Language Models
A Bayesian framework disentangles topic, agreement, and anchoring biases from interaction effects in LLM multi-turn dialogues, revealing convergence to attractors that shift with fine-tuning.
-
Informing AI Policy Assessment using Large-Scale Simulation of Interventions
A genetic algorithm exploring billions of policy combinations, scored by LLM-evaluated harm mitigation, expert cost, and participatory ratings, identifies viable AI policy options under different weighting schemes.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.