Black-Box Model Auditing

Follow this topic Ask about this topic

Momentum

12 papers in the last four weeks, up 100% on the four weeks before. 0.1% of all new papers.

Jul 13Week of Sep 28

Latest papers 66

All topics

Detecting Data Poisoning in Code Generation LLMs via Black-Box, Vulnerability-Oriented Scanning
Mar 17, 2026Shenao Yan, Shan Jin, Shimaa Ahmed +4LLM Auditing Backdoor Detection
University of Connecticut · Visa Research
Real Money, Fake Models: Deceptive Model Claims in Shadow APIs
Mar 2, 2026Yage Zhang, Yukun Jiang, Zeyuan Chen +3LLM Auditing Black-Box Model Auditing
CISPA Helmholtz Center for Information Security
Audit Me If You Can: Query-Efficient Active Fairness Auditing of Black-Box LLMs
Jan 6, 2026David Hartmann, Lena Pohlmann, Lelia Hanslik +3LLM Auditing Algorithmic Fairness
Weizenbaum Institut Berlin · Technische Universität Berlin · FIZ Karlsruhe +2
Data Provenance Auditing of Fine-Tuned Large Language Models with a Text-Preserving Technique
Oct 7, 2025Yanming Li, Cédric Eichler, Nicolas Anciaux +3Data Provenance Dataset Watermarking
PETSCRAFT · Petscraft, Inria, Palaiseau, France · INSA CVL, Bourges, France +6
Optimizing Canaries for Privacy Auditing with Metagradient Descent
Jul 21, 2025Matteo Boglioni, Terrance Liu, Andrew Ilyas +1Privacy Auditing Differentially Private Stochastic Gradient Descent
Carnegie Mellon University · Stanford Statistics
CBW: Towards Dataset Ownership Verification for Speaker Verification via Clustering-based Backdoor Watermarking
Mar 2, 2025Yiming Li, Kaiying Yan, Jiawen Diao +4Speaker Verification Dataset Watermarking
College of Computing and Data Science, Nanyang Technological University, Singapore, 639798, Singapore · School of Artificial Intelligence, Shanghai Jiao Tong University, Shanghai, 200030, China · Beijing Electronic Science and Technology Institute, Beijing, 100070, China +4

Mar 17, 2026, cs.CRJ/K move · Enter open · S save

Detecting Data Poisoning in Code Generation LLMs via Black-Box, Vulnerability-Oriented Scanning

Shenao Yan, Shan Jin, Shimaa Ahmed, Sunpreet Singh Arora+3

University of Connecticut · Visa Research

LLM Auditing Backdoor Detection Data Poisoning Attacks Black-Box Model Auditing

Open paper AskPDF arXiv

Code generation large language models (LLMs) are increasingly integrated into modern software development workflows. Recent work has shown that these models are vulnerable to backdoor and poisoning attacks that induce the generation of insecure code, yet effective defenses remain limited. Existing scanning approaches rely on token-level generation consistency to invert attack targets, which is ineffective for source code where identical semantics can appear in diverse syntactic forms. We present CodeScan, the first black-box, vulnerability-specific scanning framework for auditing code generation LLMs, assuming that the defender specifies the target vulnerability classes and provides corresponding task-relevant prompts. CodeScan identifies attack targets by analyzing structural similarities across multiple generations conditioned on different clean prompts. It combines iterative divergence analysis with abstract syntax tree (AST)-based normalization to abstract away surface-level variation and unify semantically equivalent code, isolating structures that recur consistently across generations. CodeScan then applies LLM-based vulnerability analysis to determine whether the extracted structures contain security vulnerabilities and flags the model as compromised when such a structure is found. We evaluate CodeScan against four representative attacks under both backdoor and poisoning settings across three real-world vulnerability classes. Experiments on 117 models spanning three architectures and multiple model sizes demonstrate 97%+ detection accuracy with substantially lower false positives than prior methods.

Momentum

Related topics

Latest papers 66

Detecting Data Poisoning in Code Generation LLMs via Black-Box, Vulnerability-Oriented Scanning

Real Money, Fake Models: Deceptive Model Claims in Shadow APIs

Audit Me If You Can: Query-Efficient Active Fairness Auditing of Black-Box LLMs

Data Provenance Auditing of Fine-Tuned Large Language Models with a Text-Preserving Technique

Optimizing Canaries for Privacy Auditing with Metagradient Descent

CBW: Towards Dataset Ownership Verification for Speaker Verification via Clustering-based Backdoor Watermarking