wovepaper — the arXiv observatory

papers

Publications (37)

cs.AI2026

Building Coding Agents via Entropy-Enhanced Multi-Turn Preference Optimization

Jiahao Yu, Zelei Cheng, Xian Wu +1

cs.CR2024

Take a Step Further: Understanding Page Spray in Linux Kernel Exploitation

Ziyi Guo, Dang K Le, Zhenpeng Lin +6

cs.CL2026

Decoupled Alignment for Robust Plug-and-Play Adaptation

Haozheng Luo, Jiahao Yu, Wenxin Zhang +9

The paper proposes a training-free, plug-and-play method that uses knowledge distillation and model fusion to correct misaligned (shadow-aligned) large language models, improving s…

Publications (37)

Building Coding Agents via Entropy-Enhanced Multi-Turn Preference Optimization

Take a Step Further: Understanding Page Spray in Linux Kernel Exploitation

Decoupled Alignment for Robust Plug-and-Play Adaptation

Assessing Prompt Injection Risks in 200+ Custom GPTs

Intrusions into Privacy in Video Chat Environments: Attacks and Countermeasures

Fine-Grained Library Customization

BACKDOORL: Backdoor Attack against Competitive Reinforcement Learning

From Physical to Cyber: Escalating Protection for Personalized Auto Insurance

Patch Validation in Automated Vulnerability Repair

BlockScan: Detecting Anomalies in Blockchain Transactions

A Survey on Explainable Deep Reinforcement Learning

Soft-Label Integration for Robust Toxicity Classification

UC-MOA: Utility-Conditioned Multi-Objective Alignment for Distributional Pareto-Optimality

Learning Adversary-Resistant Deep Neural Networks

An Empirical Study of Spam and Prevention Mechanisms in Online Video Chat Services

Wrinkling of Randomly Heterogeneous Film-Substrate Systems

ShadowBound: Efficient Heap Memory Protection Through Advanced Metadata Management and Customized Compiler Optimization

GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Prompts

Explaining Deep Learning Models - A Bayesian Non-parametric Approach

GPO: Learning from Critical Steps to Improve LLM Reasoning

TrustShadow: Secure Execution of Unmodified Applications with ARM TrustZone

Using Non-invertible Data Transformations to Build Adversarial-Robust Neural Networks

SafeVchat: Detecting Obscene Content and Misbehaving Users in Online Video Chat Services

Mind the Inconspicuous: Revealing the Hidden Weakness in Aligned LLMs' Refusal Boundaries

RICE: Breaking Through the Training Bottlenecks of Reinforcement Learning with Explanation

CAMP: Compiler and Allocator-based Heap Memory Protection

Adversary Resistant Deep Neural Networks with an Application to Malware Detection

TABOR: A Highly Accurate Approach to Inspecting and Restoring Trojan Backdoors in AI Systems

A Comparative Study of Rule Extraction for Recurrent Neural Networks

Towards Interrogating Discriminative Machine Learning Models

DANCE: Enhancing saliency maps using decoys

Preventing Timing Side-Channels via Security-Aware Just-In-Time Compilation

PTrix: Efficient Hardware-Assisted Fuzzing for COTS Binary

BandFuzz: An ML-powered Collaborative Fuzzing Framework

Context-aware System Service Call-oriented Symbolic Execution of Android Framework with Application to Exploit Generation

An Empirical Evaluation of Rule Extraction from Recurrent Neural Networks

Harnessing Large Language Models for Seed Generation in Greybox Fuzzing