Implicit Bias-Like Patterns in Reasoning Models

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Lee, Messi H. J., Lai, Calvin K.
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911568029024256
author Lee, Messi H. J.
Lai, Calvin K.
author_facet Lee, Messi H. J.
Lai, Calvin K.
contents Implicit biases refer to automatic mental processes that shape perceptions, judgments, and behaviors. Previous research on "implicit bias" in LLMs focused primarily on outputs rather than the processes underlying the outputs. We present the Reasoning Model Implicit Association Test (RM-IAT) to study implicit bias-like processing in reasoning models, LLMs that use step-by-step reasoning to solve complex tasks. Using RM-IAT, we find that reasoning models like o3-mini, DeepSeek-R1, gpt-oss-20b, and Qwen-3 8B consistently expend more reasoning tokens on association-incompatible tasks than association-compatible tasks, suggesting greater computational effort when processing counter-stereotypical information. Conversely, Claude 3.7 Sonnet exhibited reversed patterns, which thematic analysis associated with its unique internal focus on reasoning about bias and stereotypes. These findings demonstrate that reasoning models exhibit distinct implicit bias-like patterns and that these patterns vary significantly depending on the models' internal reasoning content.
format Preprint
id arxiv_https___arxiv_org_abs_2503_11572
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Implicit Bias-Like Patterns in Reasoning Models
Lee, Messi H. J.
Lai, Calvin K.
Computers and Society
Artificial Intelligence
Implicit biases refer to automatic mental processes that shape perceptions, judgments, and behaviors. Previous research on "implicit bias" in LLMs focused primarily on outputs rather than the processes underlying the outputs. We present the Reasoning Model Implicit Association Test (RM-IAT) to study implicit bias-like processing in reasoning models, LLMs that use step-by-step reasoning to solve complex tasks. Using RM-IAT, we find that reasoning models like o3-mini, DeepSeek-R1, gpt-oss-20b, and Qwen-3 8B consistently expend more reasoning tokens on association-incompatible tasks than association-compatible tasks, suggesting greater computational effort when processing counter-stereotypical information. Conversely, Claude 3.7 Sonnet exhibited reversed patterns, which thematic analysis associated with its unique internal focus on reasoning about bias and stereotypes. These findings demonstrate that reasoning models exhibit distinct implicit bias-like patterns and that these patterns vary significantly depending on the models' internal reasoning content.
title Implicit Bias-Like Patterns in Reasoning Models
topic Computers and Society
Artificial Intelligence
url https://arxiv.org/abs/2503.11572