Cover Image

Hidden Thoughts Are Not Secret: Reasoning Trace Exposure in LLMs

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Lu, Yu-An, Tsai, Ci-Yang, Tsai, Yu-Lin, Popa, Raluca Ada, Yu, Chia-Mu
Format:	Preprint
Published:	2026
Subjects:	Artificial Intelligence Cryptography and Security
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Opal: Private Memory for Personal AI
by: Kaviani, Darya, et al.
Published: (2026)

A Framework for Evaluating Emerging Cyberattack Capabilities of AI
by: Rodriguez, Mikel, et al.
Published: (2025)

Onyx: Cost-Efficient Disk-Oblivious ANN Search
by: Rathee, Deevashwer, et al.
Published: (2026)

Concept-Aware Privacy Mechanisms for Defending Embedding Inversion Attacks
by: Tsai, Yu-Che, et al.
Published: (2026)

MiniScope: A Least Privilege Framework for Authorizing Tool Calling Agents
by: Zhu, Jinhao, et al.
Published: (2025)

Differentially Private Fine-Tuning of Diffusion Models
by: Tsai, Yu-Lin, et al.
Published: (2024)

One Leak Away: How Pretrained Model Exposure Amplifies Jailbreak Risks in Finetuned LLMs
by: Tan, Yixin, et al.
Published: (2025)

BadThink: Triggered Overthinking Attacks on Chain-of-Thought Reasoning in Large Language Models
by: Liu, Shuaitong, et al.
Published: (2025)

SUDP: Secret-Use Delegation Protocol for Agentic Systems
by: Yu, Xiaohang, et al.
Published: (2026)

ExplainableGuard: Interpretable Adversarial Defense for Large Language Models Using Chain-of-Thought Reasoning
by: Guan, Shaowei, et al.
Published: (2025)

Understanding Secret Leakage Risks in Code LLMs: A Tokenization Perspective
by: Chen, Meifang, et al.
Published: (2026)

SFCoT: Safer Chain-of-Thought via Active Safety Evaluation and Calibration
by: Pan, Yu, et al.
Published: (2026)

MPC-Minimized Secure LLM Inference
by: Rathee, Deevashwer, et al.
Published: (2024)

Strengthening Human-Centric Chain-of-Thought Reasoning Integrity in LLMs via a Structured Prompt Framework
by: Zhou, Jiling, et al.
Published: (2026)

Auditing Black-Box LLM APIs with a Rank-Based Uniformity Test
by: Zhu, Xiaoyuan, et al.
Published: (2025)

Red Teaming Large Reasoning Models
by: Chen, Jiawei, et al.
Published: (2025)

Security and Resilience in Autonomous Vehicles: A Proactive Design Approach
by: Tsai, Chieh, et al.
Published: (2026)

Exposing Hidden Interfaces: LLM-Guided Type Inference for Reverse Engineering macOS Private Frameworks
by: Kharlamova, Arina, et al.
Published: (2026)

Multi-Designated Detector Watermarking for Language Models
by: Huang, Zhengan, et al.
Published: (2024)

Can Reasoning Models Obfuscate Reasoning? Stress-Testing Chain-of-Thought Monitorability
by: Zolkowski, Artur, et al.
Published: (2025)

MaXsive: High-Capacity and Robust Training-Free Generative Image Watermarking in Diffusion Models
by: Mao, Po-Yuan, et al.
Published: (2025)

Protect Your Secrets: Understanding and Measuring Data Exposure in VSCode Extensions
by: Liu, Yue, et al.
Published: (2024)

DiffuseTrace: A Transparent and Flexible Watermarking Scheme for Latent Diffusion Model
by: Lei, Liangqi, et al.
Published: (2024)

Web Agents Should Adopt the Plan-Then-Execute Paradigm
by: Piet, Julien, et al.
Published: (2026)

LLMs Can Covertly Sandbag on Capability Evaluations Against Chain-of-Thought Monitoring
by: Li, Chloe, et al.
Published: (2025)

Reasoning Under Pressure: How do Training Incentives Influence Chain-of-Thought Monitorability?
by: MacDermott, Matt, et al.
Published: (2025)

Rethinking Reasoning: A Survey on Reasoning-based Backdoors in LLMs
by: Hu, Man, et al.
Published: (2025)

PROMPTFUZZ: Harnessing Fuzzing Techniques for Robust Testing of Prompt Injection in LLMs
by: Yu, Jiahao, et al.
Published: (2024)

Committed SAE-Feature Traces for Audited-Session Substitution Detection in Hosted LLMs
by: Liu, Ziyang
Published: (2026)

Token-Efficient Prompt Injection Attack: Provoking Cessation in LLM Reasoning via Adaptive Token Compression
by: Cui, Yu, et al.
Published: (2025)

DPrivBench: Benchmarking LLMs' Reasoning for Differential Privacy
by: Wang, Erchi, et al.
Published: (2026)

Why Do Aligned LLMs Remain Jailbreakable: Refusal-Escape Directions, Operator-Level Sources, and Safety-Utility Trade-off
by: Chen, Yu, et al.
Published: (2026)

CachePrune: Teaching LLMs What Not to Follow via KV-Cache Editing
by: Wang, Rui, et al.
Published: (2025)

Engineering Robustness into Personal Agents with the AI Workflow Store
by: Geambasu, Roxana, et al.
Published: (2026)

Enhancing Reasoning Capacity of SLM using Cognitive Enhancement
by: Pan, Jonathan, et al.
Published: (2024)

Shattering the Echo Chamber: Hidden Safeguards in Manuscripts Against the AI Takeover of Peer Review
by: Ma, Oubo, et al.
Published: (2026)

PPMI: Privacy-Preserving LLM Interaction with Socratic Chain-of-Thought Reasoning and Homomorphically Encrypted Vector Databases
by: Bae, Yubeen, et al.
Published: (2025)

Tricking LLM-Based NPCs into Spilling Secrets
by: Shiomi, Kyohei, et al.
Published: (2025)

On the Feasibility of Using MultiModal LLMs to Execute AR Social Engineering Attacks
by: Bi, Ting, et al.
Published: (2025)

Confusion is the Final Barrier: Rethinking Jailbreak Evaluation and Investigating the Real Misuse Threat of LLMs
by: Yan, Yu, et al.
Published: (2025)