Enregistré dans:
Détails bibliographiques
Auteurs principaux: Li, Haoyang, Chen, Xuejia, XU, Zhanchao, Li, Darian, Hu, Nicole, Teng, Fei, Li, Yiming, Qiu, Luyu, Zhang, Chen Jason, Li, Qing, Chen, Lei
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2502.11075
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866908390481985536
author Li, Haoyang
Chen, Xuejia
XU, Zhanchao
Li, Darian
Hu, Nicole
Teng, Fei
Li, Yiming
Qiu, Luyu
Zhang, Chen Jason
Li, Qing
Chen, Lei
author_facet Li, Haoyang
Chen, Xuejia
XU, Zhanchao
Li, Darian
Hu, Nicole
Teng, Fei
Li, Yiming
Qiu, Luyu
Zhang, Chen Jason
Li, Qing
Chen, Lei
contents Large Language Models (LLMs) have demonstrated impressive capabilities in natural language processing tasks, such as text generation and semantic understanding. However, their performance on numerical reasoning tasks, such as basic arithmetic, numerical retrieval, and magnitude comparison, remains surprisingly poor. This gap arises from their reliance on surface-level statistical patterns rather than understanding numbers as continuous magnitudes. Existing benchmarks primarily focus on either linguistic competence or structured mathematical problem-solving, neglecting fundamental numerical reasoning required in real-world scenarios. To bridge this gap, we propose NumericBench, a comprehensive benchmark to evaluate six fundamental numerical capabilities: number recognition, arithmetic operations, contextual retrieval, comparison, summary, and logical reasoning. NumericBench includes datasets ranging from synthetic number lists to the crawled real-world data, addressing challenges like long contexts, noise, and multi-step reasoning. Extensive experiments on state-of-the-art LLMs, including GPT-4 and DeepSeek, reveal persistent weaknesses in numerical reasoning, highlighting the urgent need to improve numerically-aware language modeling. The benchmark is released in: https://github.com/TreeAI-Lab/NumericBench.
format Preprint
id arxiv_https___arxiv_org_abs_2502_11075
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Exposing Numeracy Gaps: A Benchmark to Evaluate Fundamental Numerical Abilities in Large Language Models
Li, Haoyang
Chen, Xuejia
XU, Zhanchao
Li, Darian
Hu, Nicole
Teng, Fei
Li, Yiming
Qiu, Luyu
Zhang, Chen Jason
Li, Qing
Chen, Lei
Computation and Language
Artificial Intelligence
Large Language Models (LLMs) have demonstrated impressive capabilities in natural language processing tasks, such as text generation and semantic understanding. However, their performance on numerical reasoning tasks, such as basic arithmetic, numerical retrieval, and magnitude comparison, remains surprisingly poor. This gap arises from their reliance on surface-level statistical patterns rather than understanding numbers as continuous magnitudes. Existing benchmarks primarily focus on either linguistic competence or structured mathematical problem-solving, neglecting fundamental numerical reasoning required in real-world scenarios. To bridge this gap, we propose NumericBench, a comprehensive benchmark to evaluate six fundamental numerical capabilities: number recognition, arithmetic operations, contextual retrieval, comparison, summary, and logical reasoning. NumericBench includes datasets ranging from synthetic number lists to the crawled real-world data, addressing challenges like long contexts, noise, and multi-step reasoning. Extensive experiments on state-of-the-art LLMs, including GPT-4 and DeepSeek, reveal persistent weaknesses in numerical reasoning, highlighting the urgent need to improve numerically-aware language modeling. The benchmark is released in: https://github.com/TreeAI-Lab/NumericBench.
title Exposing Numeracy Gaps: A Benchmark to Evaluate Fundamental Numerical Abilities in Large Language Models
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2502.11075