Metraipolis › archive › dc_892c04091cb9
Beyond Raw Transcripts: Structured Persona Extraction for LLM-Based Digital Twins — arXiv:2608.20344
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.20344
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T10:18:03.313Z
- SHA-256 of the archived text
cafdb4b9139de266ce61f5dc7599b5bfcf40ae7cc0c120a5cc30a115dc886cfe- Stable URL of this page
https://metraipolis.com/documento/dc_892c04091cb9
How to cite
Beyond Raw Transcripts: Structured Persona Extraction for LLM-Based Digital Twins — arXiv:2608.20344 — arXiv (en), https://arxiv.org/abs/2608.20344 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_892c04091cb9 (Metraipolis archive dc_892c04091cb9, sha256 cafdb4b9139de266ce61f5dc7599b5bfcf40ae7cc0c120a5cc30a115dc886cfe)
BibTeX
@misc{dc892c04091cb9,
title = {Beyond Raw Transcripts: Structured Persona Extraction for LLM-Based Digital Twins — arXiv:2608.20344},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.20344},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_892c04091cb9; sha256 cafdb4b9139de266ce61f5dc7599b5bfcf40ae7cc0c120a5cc30a115dc886cfe},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.20344. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
LLM-based "digital twins" aim to simulate how an individual would behavein new environments or respond to novel questions, given some representation of that individual's prior responses. A common approach constructs this representation from survey transcripts or summaries responses. Prior work shows that compressing long transcripts into shorter LLM-generated summaries does not significantly reduce predictive accuracy, suggesting that information volume is not the primary bottleneck. In this work, we argue that the key limitation is instead structural:how persona information is organized before being provided to thesimulator model. We study this by comparing unstructured summaries with structured persona representations. First, we introduce a hand-craftedschema (BDE: Background, Decision procedure, Evaluation), grounded in consumer-behavior theory, and show that it improves predictive accuracy over raw transcripts by +1.91 percentage points on a homogeneous benchmark (Twin-2K-500), with similar gains on gpt-5.4-mini and Qwen3-8B as robustness checks. However, this fixed structure does not generalizeacross more heterogeneous tasks, where performance is statistically indistinguishable from the raw transcript baseline. To address this limitation, we propose an automatic structure-discovery pipeline in which an LLM iteratively proposes and refines task-specific persona structures and extraction prompts. On a benchmark of 13 diverse sub-studies, this approach restores performance, improving mean accuracy by +1.91 percentage points over the raw transcript baseline and eliminating significant losses observed with the fixed schema. Overall, our results suggest that the main constraint in LLM-based digital twins is not how much information is provided, but how it is structured -- and that the optimal structure depends on the task. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-24 in cs.CL). Categorie arXiv: Computation and Language (cs.CL) ; Computers and Society (cs.CY). Pagina del lavoro: https://arxiv.org/abs/2608.20344 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-18T10:17:37.539Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-24?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).