Breaking the Assistant Mold: Modeling Behavioral Variation in LLM Based Procedural Character Generation

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Qraitem, Maan, Saenko, Kate, Plummer, Bryan A.
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866914527976620032
author Qraitem, Maan
Saenko, Kate
Plummer, Bryan A.
author_facet Qraitem, Maan
Saenko, Kate
Plummer, Bryan A.
contents Procedural content generation has enabled vast virtual worlds through levels, maps, and quests, but large-scale character generation remains underexplored. We identify two alignment-induced biases in existing methods: a positive moral bias, where characters uniformly adopt agreeable stances (e.g. always saying lying is bad), and a helpful assistant bias, where characters invariably answer questions directly (e.g. never refusing or deflecting). While such tendencies suit instruction-following systems, they suppress dramatic tension and yield predictable characters, stemming from maximum likelihood training and assistant fine-tuning. To address this, we introduce PersonaWeaver, a framework that disentangles world-building (roles, demographics) from behavioral-building (moral stances, interactional styles), yielding characters with more diverse reactions and moral stances, as well as second-order diversity in stylistic markers like length, tone, and punctuation. Code: https://github.com/mqraitem/Persona-Weaver
format Preprint
id arxiv_https___arxiv_org_abs_2601_03396
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Breaking the Assistant Mold: Modeling Behavioral Variation in LLM Based Procedural Character Generation
Qraitem, Maan
Saenko, Kate
Plummer, Bryan A.
Computation and Language
Procedural content generation has enabled vast virtual worlds through levels, maps, and quests, but large-scale character generation remains underexplored. We identify two alignment-induced biases in existing methods: a positive moral bias, where characters uniformly adopt agreeable stances (e.g. always saying lying is bad), and a helpful assistant bias, where characters invariably answer questions directly (e.g. never refusing or deflecting). While such tendencies suit instruction-following systems, they suppress dramatic tension and yield predictable characters, stemming from maximum likelihood training and assistant fine-tuning. To address this, we introduce PersonaWeaver, a framework that disentangles world-building (roles, demographics) from behavioral-building (moral stances, interactional styles), yielding characters with more diverse reactions and moral stances, as well as second-order diversity in stylistic markers like length, tone, and punctuation. Code: https://github.com/mqraitem/Persona-Weaver
title Breaking the Assistant Mold: Modeling Behavioral Variation in LLM Based Procedural Character Generation
topic Computation and Language
url https://arxiv.org/abs/2601.03396