Cover Image

LM Agents May Fail to Act on Their Own Risk Knowledge

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Tang, Yuzhi, Li, Tianxiao, Li, Elizabeth, Maddison, Chris J., Dong, Honghua, Ruan, Yangjun
Format:	Preprint
Published:	2025
Subjects:	Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Identifying the Risks of LM Agents with an LM-Emulated Sandbox
by: Ruan, Yangjun, et al.
Published: (2023)

Observational Scaling Laws and the Predictability of Language Model Performance
by: Ruan, Yangjun, et al.
Published: (2024)

APPL: A Prompt Programming Language for Harmonious Integration of Programs and Large Language Model Prompts
by: Dong, Honghua, et al.
Published: (2024)

Reasoning to Learn from Latent Thoughts
by: Ruan, Yangjun, et al.
Published: (2025)

Formally Solving Answer-Construction Problems in Lean
by: Sun, Jialiang, et al.
Published: (2025)

Test-Time Fairness and Robustness in Large Language Models
by: Cotta, Leonardo, et al.
Published: (2024)

Agent Benchmarks Fail Public Sector Requirements
by: Rystrøm, Jonathan, et al.
Published: (2026)

When Agents Fail to Act: A Diagnostic Framework for Tool Invocation Reliability in Multi-Agent LLM Systems
by: Huang, Donghao, et al.
Published: (2026)

When Single-Agent with Skills Replace Multi-Agent Systems and When They Fail
by: Li, Xiaoxiao
Published: (2026)

WebSuite: Systematically Evaluating Why Web Agents Fail
by: Li, Eric, et al.
Published: (2024)

$τ^2$-Bench: Evaluating Conversational Agents in a Dual-Control Environment
by: Barres, Victor, et al.
Published: (2025)

AgentFly: Extensible and Scalable Reinforcement Learning for LM Agents
by: Wang, Renxi, et al.
Published: (2025)

A Survey of Generative AI for de novo Drug Design: New Frontiers in Molecule and Protein Generation
by: Tang, Xiangru, et al.
Published: (2024)

KG-BiLM: Knowledge Graph Embedding via Bidirectional Language Models
by: Chen, Zirui, et al.
Published: (2025)

Why Do Multi-Agent LLM Systems Fail?
by: Cemri, Mert, et al.
Published: (2025)

Where LLM Agents Fail and How They can Learn From Failures
by: Zhu, Kunlun, et al.
Published: (2025)

Standard Benchmarks Fail -- Auditing LLM Agents in Finance Must Prioritize Risk
by: Chen, Zichen, et al.
Published: (2025)

A Multi-LLM-Agent-Based Framework for Economic and Public Policy Analysis
by: Hao, Yuzhi, et al.
Published: (2025)

SememeLM: A Sememe Knowledge Enhanced Method for Long-tail Relation Representation
by: Li, Shuyi, et al.
Published: (2024)

Co-ReAct: Rubrics as Step-Level Collaborators for ReAct Agents
by: Kang, Jiazheng, et al.
Published: (2026)

Exploring the Role of Knowledge Graph-Based RAG in Japanese Medical Question Answering with Small-Scale LLMs
by: Chen, Yingjian, et al.
Published: (2025)

Exploring Autonomous Agents: A Closer Look at Why They Fail When Completing Tasks
by: Lu, Ruofan, et al.
Published: (2025)

Graph-based Uncertainty Metrics for Long-form Language Model Outputs
by: Jiang, Mingjian, et al.
Published: (2024)

Why Retrying Fails: Context Contamination in LLM Agent Pipelines
by: Yang, Zhanfu
Published: (2026)

How Coding Agents Fail Their Users: A Large-Scale Analysis of Developer-Agent Misalignment in 20,574 Real-World Sessions
by: Tang, Ningzhi, et al.
Published: (2026)

ReaLM: Residual Quantization Bridging Knowledge Graph Embeddings and Large Language Models
by: Guo, Wenbin, et al.
Published: (2025)

PreAct: Prediction Enhances Agent's Planning Ability
by: Fu, Dayuan, et al.
Published: (2024)

PoAct: Policy and Action Dual-Control Agent for Generalized Applications
by: Yuan, Guozhi, et al.
Published: (2025)

Language Models Fail to Introspect About Their Knowledge of Language
by: Song, Siyuan, et al.
Published: (2025)

Current Agents Fail to Leverage World Model as Tool for Foresight
by: Qian, Cheng, et al.
Published: (2026)

The Reasons that Agents Act: Intention and Instrumental Goals
by: Ward, Francis Rhys, et al.
Published: (2024)

Schema-Aware Multi-Task Learning for Complex Text-to-SQL
by: Wu, Yangjun, et al.
Published: (2024)

ExplicitLM: Decoupling Knowledge from Parameters via Explicit Memory Banks
by: Yu, Chengzhang, et al.
Published: (2025)

Focused ReAct: Improving ReAct through Reiterate and Early Stop
by: Li, Shuoqiu, et al.
Published: (2024)

Your Agent May Misevolve: Emergent Risks in Self-evolving LLM Agents
by: Shao, Shuai, et al.
Published: (2025)

LM Agents for Coordinating Multi-User Information Gathering
by: Jhamtani, Harsh, et al.
Published: (2025)

Micro-Act: Mitigating Knowledge Conflict in LLM-based RAG via Actionable Self-Reasoning
by: Huo, Nan, et al.
Published: (2025)

Butterfly Effects in Toolchains: A Comprehensive Analysis of Failed Parameter Filling in LLM Tool-Agent Systems
by: Xiong, Qian, et al.
Published: (2025)

ReAct Meets ActRe: When Language Agents Enjoy Training Data Autonomy
by: Yang, Zonghan, et al.
Published: (2024)

GroundAct: Can LLM Agents Ground Actions in Environmental States?
by: Wang, Zixuan, et al.
Published: (2025)