A Framework for Benchmarking and Aligning Task-Planning Safety in LLM-Based Embodied Agents

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Huang, Yuting, Ding, Leilei, Tang, Zhipeng, Wang, Tianfu, Lin, Xinrui, Zhang, Wuyang, Ma, Mingxiao, Zhang, Yanyong
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910915277881344
author Huang, Yuting
Ding, Leilei
Tang, Zhipeng
Wang, Tianfu
Lin, Xinrui
Zhang, Wuyang
Ma, Mingxiao
Zhang, Yanyong
author_facet Huang, Yuting
Ding, Leilei
Tang, Zhipeng
Wang, Tianfu
Lin, Xinrui
Zhang, Wuyang
Ma, Mingxiao
Zhang, Yanyong
contents Large Language Models (LLMs) exhibit substantial promise in enhancing task-planning capabilities within embodied agents due to their advanced reasoning and comprehension. However, the systemic safety of these agents remains an underexplored frontier. In this study, we present Safe-BeAl, an integrated framework for the measurement (SafePlan-Bench) and alignment (Safe-Align) of LLM-based embodied agents' behaviors. SafePlan-Bench establishes a comprehensive benchmark for evaluating task-planning safety, encompassing 2,027 daily tasks and corresponding environments distributed across 8 distinct hazard categories (e.g., Fire Hazard). Our empirical analysis reveals that even in the absence of adversarial inputs or malicious intent, LLM-based agents can exhibit unsafe behaviors. To mitigate these hazards, we propose Safe-Align, a method designed to integrate physical-world safety knowledge into LLM-based embodied agents while maintaining task-specific performance. Experiments across a variety of settings demonstrate that Safe-BeAl provides comprehensive safety validation, improving safety by 8.55 - 15.22%, compared to embodied agents based on GPT-4, while ensuring successful task completion.
format Preprint
id arxiv_https___arxiv_org_abs_2504_14650
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A Framework for Benchmarking and Aligning Task-Planning Safety in LLM-Based Embodied Agents
Huang, Yuting
Ding, Leilei
Tang, Zhipeng
Wang, Tianfu
Lin, Xinrui
Zhang, Wuyang
Ma, Mingxiao
Zhang, Yanyong
Artificial Intelligence
Large Language Models (LLMs) exhibit substantial promise in enhancing task-planning capabilities within embodied agents due to their advanced reasoning and comprehension. However, the systemic safety of these agents remains an underexplored frontier. In this study, we present Safe-BeAl, an integrated framework for the measurement (SafePlan-Bench) and alignment (Safe-Align) of LLM-based embodied agents' behaviors. SafePlan-Bench establishes a comprehensive benchmark for evaluating task-planning safety, encompassing 2,027 daily tasks and corresponding environments distributed across 8 distinct hazard categories (e.g., Fire Hazard). Our empirical analysis reveals that even in the absence of adversarial inputs or malicious intent, LLM-based agents can exhibit unsafe behaviors. To mitigate these hazards, we propose Safe-Align, a method designed to integrate physical-world safety knowledge into LLM-based embodied agents while maintaining task-specific performance. Experiments across a variety of settings demonstrate that Safe-BeAl provides comprehensive safety validation, improving safety by 8.55 - 15.22%, compared to embodied agents based on GPT-4, while ensuring successful task completion.
title A Framework for Benchmarking and Aligning Task-Planning Safety in LLM-Based Embodied Agents
topic Artificial Intelligence
url https://arxiv.org/abs/2504.14650