4 citations · 4 across the 18 of their papers we have counts for
27 papers
A Low Similarity Score Does Not Show Independence after Source-Conditioned Repository Rewriting
Micheline Bénédicte Moumoula, Serge Lionel Nikiema, Tailia Malloy +3
Maintainers sometimes cite a low token-similarity score as evidence that an LLM-assisted rewrite of a copyleft project is independent of the original. We test what such a score can…
RiskyPy: Toward Model-Calibrated Prompt Generation for Joint Security and Functionality Evaluation of LLM-Generated Code
Zacharie Chenail-Larcher, Daniele Lunghi, Melissa Tessa +5
Evaluating approaches that improve the security of LLM code generation requires datasets that jointly assess security and functionality and remain adapted to the capabilities of th…
Program Bindings Inside Code Models
Prateek Kumar Rajput, Yewei Song, Jacques Klein +1
Code language models complete, review, and repair programs without exposing the program structure encoded in their hidden states. In this paper, we test whether the hidden states o…
Metacognitive Reasoning in Energy Based Models using Instance Based Learning Theory
Tailia Malloy, Prateek Kumar Rajput, Serge Lionel Nikiema +2
Metacognition involves reasoning about cognitive processes themselves. An example is in resource allocation where we choose how much time and effort to put into a reasoning task be…
Trajectory-Level Security Debt in LLM Coding Agents
Prateek Kumar Rajput, Abdoul Kader Kabore, Yewei Song +4
LLM coding agents can traverse hundreds of intermediate code states before submitting a solution. Evaluating only the final artifact leaves the evolution of security findings unmea…
Right Facts, Wrong Rule: Scope-Transfer Poisoning of Retrieval-Augmented Generation
Kaisheng Fan, Yishu Gao, Xunzhu Tang +2
Retrieval-augmented generation (RAG) systems use external documents for facts and rules that guide answers. We study rule-scope misapplication, where a rule valid within its docume…
BMA: Backchain Memory Attacks Create Unauthorized Control Paths in LLM Agents
Kaisheng Fan, Yishu Gao, Xunzhu Tang +2
Persistent memory enables LLM agents to reuse prior experience, but creates a new security boundary: what an agent may remember is not what it should act on. We expose an unauthori…
Evaluating Inference-Time Defenses Against Package Hallucination in LLM-Generated Code
Alberick Euraste Djire, Iyiola E. Olatunji, Melissa Tessa +3
LLMs are increasingly used for code generation, yet they frequently hallucinate non-existent software packages, creating exploitable entry points into the software supply chain. We…
Do Not Copy/Paste: Soft Barriers for Copying in AI-Assisted Programming
Iyiola E. Olatunji, Alberick Euraste Djire, Jacques Klein +1
Copying a function from a chat window into an editor takes less than a second. For many uses of AI coding tools, that speed is the point; in settings such as programming education,…
Memorization Diagnostics for Code LLMs Should be Scale-Aware
Prateek Kumar Rajput, Abdoul Aziz Bonkoungou, Alberick Euraste Djiré +6
The extent to which large language models for code rely on memorization over genuine understanding remains highly debated. While current literature frequently reports widespread me…
Have I Seen You? Embedding Behavior Signals Synthetic Face Dataset Membership
Paweł Borsukiewicz, Daniele Lunghi, Wendkûuni C. Ouédraogo +2
Synthetic face datasets are increasingly used to reduce privacy exposure and data access constraints in biometric recognition. Yet the generators that produce these datasets are tr…
Old Tricks, New Models: How Simple Image Transformations Break Modern AI-based Content Moderation
Marco Alecci, Francesco Marchiori, Iyiola Emmanuel Olatunji +2
While automated content-moderation systems have become essential for screening harmful content at scale, conventional task-specific classifiers often provide limited policy cov- er…
A Framework for Using and Evaluating LLMs as Surrogate Experts in Security Surveys: Reliability, Bias, and Implications
Despoina Giarimpampa, Roland Meier, Tegawendé F. Bissyandé +2
Expert surveys are widely used in security research to study practitioner workows and decision-making, yet recruiting domain experts - especially in Security Operations Centres (SO…
A Single Patch Is Not Enough: Deterministic Fusion of Repair Candidates
Boyang Yang, Xiangliang Hu, Luyao Ren +4
Modern LLM coding agents are commonly evaluated using pass@k, but developers typically apply a single final patch in real-world settings. This pass@k-to-pass@1 gap is a post-genera…
Detecting Malicious Agent Skills in the Wild using Attention
Bacem Etteib, Daniele Lunghi, Tégawendé F. Bissyandé
LLM agents increasingly load skills, file-based packages of natural-language instructions written by third parties and distributed through marketplaces, that execute with the user'…
When English Isn't the Best Teacher: Source Language Effects in Cross-Lingual In-Context Learning
Fred Philippy, Siwen Guo, Jacques Klein +1
Cross-lingual transfer in multilingual NLP has been widely explored in supervised fine-tuning contexts, where factors like data availability and linguistic similarity largely deter…
Predictable Confabulations: Factual Recall by LLMs Scales with Model Size and Topic Frequency
Matthew L. Smith, Jonathan P. Shock, Samuel T. Segun +2
While scaling laws govern aggregate large language model performance, no scaling law has linked factual recall to both model size and training-data composition. We evaluated 38 mod…
Evaluation Drift in LLM Personality Induction: Are We Moving the Goalpost?
Prateek Rajput, Yewei Song, Iyiola E. Olatunji +2
Can large language models reliably express a human-like personality, or are they merely mimicking surface cues without a stable underlying profile? To investigate this, we induce p…
Learned or Memorized ? Quantifying Memorization Advantage in Code LLMs
Djiré Albérick Euraste, Kaboré Abdoul Kader, Jordan Samhi +3
The lack of transparency about code datasets used to train large language models (LLMs) makes it difficult to detect, evaluate, and mitigate data leakage. We present a perturbation…
Empirical Evaluation of PDF Parsing and Chunking for Financial Question Answering with RAG
Omar El Bachyr, Yewei Song, Saad Ezzini +5
PDF files are primarily intended for human reading rather than automated processing. In addition, the heterogeneous content of PDFs, such as text, tables, and images, poses signifi…
Defusing the Trigger: Tail-Risk-Informed Attention Rebalancing for LLM Backdoor Mitigation
Kaisheng Fan, Yishu Gao, Xunzhu Tang +2
Backdoored large language models (LLMs) exhibit attacker-specified behavior at inference time while retaining normal performance on benign inputs. Existing mitigations often requir…
Adversarial Camouflage
Paweł Borsukiewicz, Daniele Lunghi, Melissa Tessa +2
While the rapid development of facial recognition algorithms has enabled numerous beneficial applications, their widespread deployment has raised significant concerns about the ris…
Correctness isnt Efficiency: Runtime Memory Divergence in LLM-Generated Code
Prateek Rajput, Yewei Song, Abdoul Aziz Bonkoungou +4
Large language models (LLMs) can generate programs that pass unit tests, but passing tests does not guarantee reliable runtime behavior. We find that different correct solutions to…
How Secure is Secure Code Generation? Adversarial Prompts Put LLM Defenses to the Test
Melissa Tessa, Iyiola E. Olatunji, Aicha War +2
Recent secure code generation methods, using vulnerability-aware fine-tuning, prefix-tuning, and prompt optimization, claim to prevent LLMs from producing insecure code. However, t…