Cover Image

TRACES: Proactive Safety Auditing for Multi-Turn LLM Agents via Trajectory-State Modeling

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Li, Jiaqian, Li, Yanshu, Zhang, Boxuan, Tang, Ruixiang, Huang, Kuan-Hao
Format:	Preprint
Published:	2026
Subjects:	Computation and Language Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Read the Scene, Not the Script: Outcome-Aware Safety for LLMs
by: Wu, Rui, et al.
Published: (2025)

Steering Vector Fields for Context-Aware Inference-Time Control in Large Language Models
by: Li, Jiaqian, et al.
Published: (2026)

TSR: Trajectory-Search Rollouts for Multi-Turn RL of LLM Agents
by: Djuhera, Aladin, et al.
Published: (2026)

Towards Generalizable Implicit In-Context Learning with Attention Routing
by: Li, Jiaqian, et al.
Published: (2025)

AgentForesight: Online Auditing for Early Failure Prediction in Multi-Agent Systems
by: Zhang, Boxuan, et al.
Published: (2026)

RAGEN: Understanding Self-Evolution in LLM Agents via Multi-Turn Reinforcement Learning
by: Wang, Zihan, et al.
Published: (2025)

TrajAgent: An LLM-Agent Framework for Trajectory Modeling via Large-and-Small Model Collaboration
by: Du, Yuwei, et al.
Published: (2024)

AgentFold: Long-Horizon Web Agents with Proactive Context Management
by: Ye, Rui, et al.
Published: (2025)

TrustAgent: Towards Safe and Trustworthy LLM-based Agents
by: Hua, Wenyue, et al.
Published: (2024)

QET: Enhancing Quantized LLM Parameters and KV cache Compression through Element Substitution and Residual Clustering
by: Wang, Yanshu, et al.
Published: (2024)

Unsafer in Many Turns: Benchmarking and Defending Multi-Turn Safety Risks in Tool-Using Agents
by: Li, Xu, et al.
Published: (2026)

Training Proactive and Personalized LLM Agents
by: Sun, Weiwei, et al.
Published: (2025)

WebAgent-R1: Training Web Agents via End-to-End Multi-Turn Reinforcement Learning
by: Wei, Zhepei, et al.
Published: (2025)

AgentGym-RL: Training LLM Agents for Long-Horizon Decision Making through Multi-Turn Reinforcement Learning
by: Xi, Zhiheng, et al.
Published: (2025)

MAATS: A Multi-Agent Automated Translation System Based on MQM Evaluation
by: Wang, George, et al.
Published: (2025)

Helpful to a Fault: Measuring Illicit Assistance in Multi-Turn, Multilingual LLM Agents
by: Talokar, Nivya, et al.
Published: (2026)

Dive into the Agent Matrix: A Realistic Evaluation of Self-Replication Risk in LLM Agents
by: Zhang, Boxuan, et al.
Published: (2025)

When Reward Hacking Rebounds: Understanding and Mitigating It with Representation-Level Signals
by: Wu, Rui, et al.
Published: (2026)

Trial and Error: Exploration-Based Trajectory Optimization for LLM Agents
by: Song, Yifan, et al.
Published: (2024)

Attention-Aware GNN-based Input Defense against Multi-Turn LLM Jailbreak
by: Huang, Zixuan, et al.
Published: (2025)

MAVEN: Multi-Agent Verification-Elaboration Network with In-Step Epistemic Auditing
by: Yao, Yinsheng, et al.
Published: (2026)

Out-of-Domain Intent Detection Considering Multi-Turn Dialogue Contexts
by: Lang, Hao, et al.
Published: (2023)

Direct Multi-Turn Preference Optimization for Language Agents
by: Shi, Wentao, et al.
Published: (2024)

CoT-UQ: Improving Response-wise Uncertainty Quantification in LLMs with Chain-of-Thought
by: Zhang, Boxuan, et al.
Published: (2025)

Eliciting Behaviors in Multi-Turn Conversations
by: Huang, Jing, et al.
Published: (2025)

ArCHer: Training Language Model Agents via Hierarchical Multi-Turn RL
by: Zhou, Yifei, et al.
Published: (2024)

PsychePass: Calibrating LLM Therapeutic Competence via Trajectory-Anchored Tournaments
by: Chen, Zhuang, et al.
Published: (2026)

Chain of Attack: a Semantic-Driven Contextual Multi-Turn attacker for LLM
by: Yang, Xikang, et al.
Published: (2024)

Few-Step Diffusion Language Models via Trajectory Self-Distillation
by: Zhang, Tunyu, et al.
Published: (2026)

DBR: Divergence-Based Regularization for Debiasing Natural Language Understanding Models
by: Li, Zihao, et al.
Published: (2025)

Navigating the Shortcut Maze: A Comprehensive Analysis of Shortcut Learning in Text Classification by Language Models
by: Zhou, Yuqing, et al.
Published: (2024)

STeCa: Step-level Trajectory Calibration for LLM Agent Learning
by: Wang, Hanlin, et al.
Published: (2025)

Planning with Multi-Constraints via Collaborative Language Agents
by: Zhang, Cong, et al.
Published: (2024)

LLM Defenses Are Not Robust to Multi-Turn Human Jailbreaks Yet
by: Li, Nathaniel, et al.
Published: (2024)

Enhancing LLM Agent Safety via Causal Influence Prompting
by: Hahm, Dongyoon, et al.
Published: (2025)

Implicit Turn-Wise Policy Optimization for Proactive User-LLM Interaction
by: Wang, Haoyu, et al.
Published: (2026)

ReAGent: A Model-agnostic Feature Attribution Method for Generative Language Models
by: Zhao, Zhixue, et al.
Published: (2024)

FragileFlow: Spectral Control of Correct-but-Fragile Predictions for Foundation Model Robustness
by: Li, Zhuoyun, et al.
Published: (2026)

VLMGuard-R1: Proactive Safety Alignment for VLMs via Reasoning-Driven Prompt Optimization
by: Chen, Menglan, et al.
Published: (2025)

Stabilizing Off-Policy Training for Long-Horizon LLM Agent via Turn-Level Importance Sampling and Clipping-Triggered Normalization
by: Li, Chenliang, et al.
Published: (2025)