"Nothing to See Here'': Unintended Disclosure through Revision Traces of LLM Deliverables

arXiv:2609.35408 · cs.CR, cs.AI · Submitted 2026-09-28 · Read on arXiv

cs.CR, cs.AI

Submitted: 2026-09-28

Updated: 2026-09-28

Code: https://github.com/TrustAIRLab/RevLeakBench.2https:

Terminology

• Trace — 233× in this paper · explained in 2 episodes
• Benchmark — 17× in this paper · explained in 2 episodes
• In International Conference on Learning Representations — 5× in this paper
• Meeting of the Association for Computational Linguistics — 5× in this paper
• Adaptation — 3× in this paper · explained in 1 episode
• Alignment — 2× in this paper · explained in 1 episode
• Benchmarking — 2× in this paper · explained in 3 episodes
• Conference on Computer and Communications Security — 2× in this paper
• Empirical Methods in Natural Language Processing — 2× in this paper
• Large Language Model (LLM) — 2× in this paper · explained in 3 episodes
• Ablation — 1× in this paper · explained in 1 episode
• Adam — 1× in this paper · explained in 2 episodes
• Agent Skills — 1× in this paper · explained in 1 episode
• Core — 1× in this paper · explained in 2 episodes
• In-Context Learning — 1× in this paper · explained in 4 episodes
• Information Leakage — 1× in this paper · explained in 2 episodes
• Protecting the Confidentiality of Personally Identifiable Information — 1× in this paper
• Reasoning Traces — 1× in this paper · explained in 3 episodes
• Steering — 1× in this paper · explained in 2 episodes

Sources

• SALT: Steering Activations towards Leakage-free Thinking in Chain of Thought
• How Your Credentials Are Leaked by LLM Agent Skills: An Empirical Study
• Chain-of-Sanitized-Thoughts: Plugging PII Leakage in CoT of Large Reasoning Models
• To Add Is Machine, To Delete Is Human: Measuring and Mitigating Deletion Avoidance in LLM Code Editing
• Capable but Careless: Do Computer-Use Agents Follow Contextual Integrity?
• Can You Keep a Secret? Involuntary Information Leakage in Language Model Writing
• EarlySciRev: A Dataset of Early-Stage Scientific Revisions Extracted from LaTeX Writing Traces
• Hidden Thoughts Are Not Secret: Reasoning Trace Exposure in LLMs
• GDPval: Evaluating AI Model Performance on Real-World Economically Valuable Tasks
• Ignore Previous Prompt: Attack Techniques For Language Models
• Assisting in Writing Wikipedia-like Articles From Scratch with Large Language Models
• PrivacyAlign: Contextual Privacy Alignment for LLM Agents
• OverleafCopilot: Empowering Academic Writing in Overleaf with Large Language Models
• WritingBench: A Comprehensive Benchmark for Generative Writing
• TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks
• ShareChat: A Dataset of Chatbot Conversations in the Wild
• Effective Prompt Extraction from Language Models
• AgentDAM: Privacy Leakage Evaluation for Autonomous Web Agents
• Multi-axis inertial sensing with 2D matter-wave arrays

Related papers

• SoK: AI-Augmented Binary Reversing
• Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
• Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
• Efficient Fuzzy PSI under One-Sided Assumptions
• Sealing the Audit-Runtime Gap for LLM Skills
• Token Composition: A Graph Based on EVM Logs