Longer Context, Deeper Thinking: Uncovering the Role of Long-Context Ability in Reasoning

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Yang, Wang, Liu, Zirui, Jin, Hongye, Yin, Qingyu, Chaudhary, Vipin, Han, Xiaotian
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912389415305216
author Yang, Wang
Liu, Zirui
Jin, Hongye
Yin, Qingyu
Chaudhary, Vipin
Han, Xiaotian
author_facet Yang, Wang
Liu, Zirui
Jin, Hongye
Yin, Qingyu
Chaudhary, Vipin
Han, Xiaotian
contents Recent language models exhibit strong reasoning capabilities, yet the influence of long-context capacity on reasoning remains underexplored. In this work, we hypothesize that current limitations in reasoning stem, in part, from insufficient long-context capacity, motivated by empirical observations such as (1) higher context window length often leads to stronger reasoning performance, and (2) failed reasoning cases resemble failed long-context cases. To test this hypothesis, we examine whether enhancing a model's long-context ability before Supervised Fine-Tuning (SFT) leads to improved reasoning performance. Specifically, we compared models with identical architectures and fine-tuning data but varying levels of long-context capacity. Our results reveal a consistent trend: models with stronger long-context capacity achieve significantly higher accuracy on reasoning benchmarks after SFT. Notably, these gains persist even on tasks with short input lengths, indicating that long-context training offers generalizable benefits for reasoning performance. These findings suggest that long-context modeling is not just essential for processing lengthy inputs, but also serves as a critical foundation for reasoning. We advocate for treating long-context capacity as a first-class objective in the design of future language models.
format Preprint
id arxiv_https___arxiv_org_abs_2505_17315
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Longer Context, Deeper Thinking: Uncovering the Role of Long-Context Ability in Reasoning
Yang, Wang
Liu, Zirui
Jin, Hongye
Yin, Qingyu
Chaudhary, Vipin
Han, Xiaotian
Artificial Intelligence
Computation and Language
Machine Learning
Recent language models exhibit strong reasoning capabilities, yet the influence of long-context capacity on reasoning remains underexplored. In this work, we hypothesize that current limitations in reasoning stem, in part, from insufficient long-context capacity, motivated by empirical observations such as (1) higher context window length often leads to stronger reasoning performance, and (2) failed reasoning cases resemble failed long-context cases. To test this hypothesis, we examine whether enhancing a model's long-context ability before Supervised Fine-Tuning (SFT) leads to improved reasoning performance. Specifically, we compared models with identical architectures and fine-tuning data but varying levels of long-context capacity. Our results reveal a consistent trend: models with stronger long-context capacity achieve significantly higher accuracy on reasoning benchmarks after SFT. Notably, these gains persist even on tasks with short input lengths, indicating that long-context training offers generalizable benefits for reasoning performance. These findings suggest that long-context modeling is not just essential for processing lengthy inputs, but also serves as a critical foundation for reasoning. We advocate for treating long-context capacity as a first-class objective in the design of future language models.
title Longer Context, Deeper Thinking: Uncovering the Role of Long-Context Ability in Reasoning
topic Artificial Intelligence
Computation and Language
Machine Learning
url https://arxiv.org/abs/2505.17315