Saved in:
Bibliographic Details
Main Authors: Wang, Yifei, Li, Tianlin, Zhang, Xiaohan, Zhang, Xiaoyu, Ma, Wei, Cheng, Mingfei, Pan, Li
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.19790
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918460961849344
author Wang, Yifei
Li, Tianlin
Zhang, Xiaohan
Zhang, Xiaoyu
Ma, Wei
Cheng, Mingfei
Pan, Li
author_facet Wang, Yifei
Li, Tianlin
Zhang, Xiaohan
Zhang, Xiaoyu
Ma, Wei
Cheng, Mingfei
Pan, Li
contents Large language models (LLMs) are increasingly deployed under diverse numerical precision configurations, including standard floating-point formats (e.g., bfloat16 and float16) and quantized integer formats (e.g., int16 and int8), to meet efficiency and resource constraints. However, minor inconsistencies between LLMs of different precisions are difficult to detect and are often overlooked by existing evaluation methods. In this paper, we present PrecisionDiff, an automated differential testing framework for systematically detecting precision-induced behavioral disagreements in LLMs. PrecisionDiff generates precision-sensitive test inputs and performs cross-precision comparative analysis to uncover subtle divergences that remain hidden under conventional testing strategies. To demonstrate its practical significance, we instantiate PrecisionDiff on the alignment verification task, where precision-induced disagreements manifest as jailbreak divergence-inputs that are rejected under one precision may produce harmful responses under another. Experimental results show that such behavioral disagreements are widespread across multiple open-source aligned LLMs and precision settings, and that PrecisionDiff significantly outperforms vanilla testing methods in detecting these issues. Our work enables automated precision-sensitive test generation, facilitating effective pre-deployment evaluation and improving precision robustness during training.
format Preprint
id arxiv_https___arxiv_org_abs_2604_19790
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Hidden Reliability Risks in Large Language Models: Systematic Identification of Precision-Induced Output Disagreements
Wang, Yifei
Li, Tianlin
Zhang, Xiaohan
Zhang, Xiaoyu
Ma, Wei
Cheng, Mingfei
Pan, Li
Artificial Intelligence
Machine Learning
I.2.7; K.6.5
Large language models (LLMs) are increasingly deployed under diverse numerical precision configurations, including standard floating-point formats (e.g., bfloat16 and float16) and quantized integer formats (e.g., int16 and int8), to meet efficiency and resource constraints. However, minor inconsistencies between LLMs of different precisions are difficult to detect and are often overlooked by existing evaluation methods. In this paper, we present PrecisionDiff, an automated differential testing framework for systematically detecting precision-induced behavioral disagreements in LLMs. PrecisionDiff generates precision-sensitive test inputs and performs cross-precision comparative analysis to uncover subtle divergences that remain hidden under conventional testing strategies. To demonstrate its practical significance, we instantiate PrecisionDiff on the alignment verification task, where precision-induced disagreements manifest as jailbreak divergence-inputs that are rejected under one precision may produce harmful responses under another. Experimental results show that such behavioral disagreements are widespread across multiple open-source aligned LLMs and precision settings, and that PrecisionDiff significantly outperforms vanilla testing methods in detecting these issues. Our work enables automated precision-sensitive test generation, facilitating effective pre-deployment evaluation and improving precision robustness during training.
title Hidden Reliability Risks in Large Language Models: Systematic Identification of Precision-Induced Output Disagreements
topic Artificial Intelligence
Machine Learning
I.2.7; K.6.5
url https://arxiv.org/abs/2604.19790