Point of Order: Action-Aware LLM Persona Modeling for Realistic Civic Simulation

TL;DR

This paper presents a pipeline to create speaker-attributed transcripts from public recordings, enabling more realistic civic simulations with large language models by modeling specific participants and their actions.

Contribution

It introduces a reproducible method to transform public recordings into detailed, speaker-attributed datasets with persona and action metadata, improving LLM civic simulation fidelity.

Findings

67% reduction in perplexity after fine-tuning

Nearly doubled classifier-based performance metrics

Human evaluations show simulations often indistinguishable from real deliberations

Abstract

Large language models offer opportunities to simulate multi-party deliberation, but realistic modeling remains limited by a lack of speaker-attributed data. Transcripts produced via automatic speech recognition (ASR) assign anonymous speaker labels (e.g., Speaker_1), preventing models from capturing consistent human behavior. This work introduces a reproducible pipeline to transform public Zoom recordings into speaker-attributed transcripts with metadata like persona profiles and pragmatic action tags (e.g., [propose_motion]). We release three local government deliberation datasets: Appellate Court hearings, School Board meetings, and Municipal Council sessions. Fine-tuning LLMs to model specific participants using this "action-aware" data produces a 67% reduction in perplexity and nearly doubles classifier-based performance metrics for speaker fidelity and realism. Turing-style human evaluations show our simulations are often indistinguishable from real deliberations, providing a practical and scalable method for complex realistic civic simulations.