21 citations · 29 across the 15 of their papers we have counts for
15 papers
Dr.LLM: Dynamic Layer Routing in LLMs
Ahmed Heakl, Martin Gubri, Salman Khan +2
Large Language Models (LLMs) process every token through all layers of a transformer stack, causing wasted computation on simple queries and insufficient flexibility for harder one…
DISCO: Diversifying Sample Condensation for Efficient Model Evaluation
Alexander Rubinstein, Benjamin Raible, Martin Gubri +1
Evaluating modern machine learning models has become prohibitively expensive. Benchmarks such as LMMs-Eval and HELM demand thousands of GPU hours per model. Costly evaluation reduc…
Leaky Thoughts: Large Reasoning Models Are Not Private Thinkers
Tommaso Green, Martin Gubri, Haritz Puerto +2
We study privacy leakage in the reasoning traces of large reasoning models used as personal agents. Unlike final outputs, reasoning traces are often assumed to be internal and safe…
C-SEO Bench: Does Conversational SEO Work?
Haritz Puerto, Martin Gubri, Tommaso Green +2
Large Language Models (LLMs) are transforming search engines into Conversational Search Engines (CSE). Consequently, Search Engine Optimization (SEO) is being shifted into Conversa…
Testing Uniform Random Samplers: Methods, Datasets and Protocols
Olivier Zeyen, Maxime Cordy, Martin Gubri +2
Boolean formulae compactly encode huge, constrained search spaces. Thus, variability-intensive systems are often encoded with Boolean formulae. The search space of a variability-in…
Social Science Is Necessary for Operationalizing Socially Responsible Foundation Models
Adam Davies, Elisa Nguyen, Michael Simeone +2
With the rise of foundation models, there is growing concern about their potential social impacts. Social science has a long history of studying the social impacts of transformativ…
Scaling Up Membership Inference: When and How Attacks Succeed on Large Language Models
Haritz Puerto, Martin Gubri, Sangdoo Yun +1
Membership inference attacks (MIA) attempt to verify the membership of a given data sample in the training set for a model. MIA has become relevant in recent years, following the r…
Calibrating Large Language Models Using Their Generations Only
Dennis Ulmer, Martin Gubri, Hwaran Lee +2
As large language models (LLMs) are increasingly deployed in user-facing applications, building trust and maintaining safety by accurately quantifying a model's confidence in its p…
TRAP: Targeted Random Adversarial Prompt Honeypot for Black-Box Identification
Martin Gubri, Dennis Ulmer, Hwaran Lee +2
Large Language Model (LLM) services and models often come with legal rules on who can use them and how they must use them. Assessing the compliance of the released LLMs is crucial,…
ProPILE: Probing Privacy Leakage in Large Language Models
Siwon Kim, Sangdoo Yun, Hwaran Lee +3
The rapid advancement and widespread use of large language models (LLMs) have raised significant concerns regarding the potential leakage of personally identifiable information (PI…
Going Further: Flatness at the Rescue of Early Stopping for Adversarial Example Transferability
Martin Gubri, Maxime Cordy, Yves Le Traon
Transferability is the property of adversarial examples to be misclassified by other models than the surrogate model for which they were crafted. Previous research has shown that e…
LGV: Boosting Adversarial Example Transferability from Large Geometric Vicinity
Martin Gubri, Maxime Cordy, Mike Papadakis +2
We propose transferability from Large Geometric Vicinity (LGV), a new technique to increase the transferability of black-box adversarial attacks. LGV starts from a pretrained surro…
Influence-Driven Data Poisoning in Graph-Based Semi-Supervised Classifiers
Adriano Franci, Maxime Cordy, Martin Gubri +2
Graph-based Semi-Supervised Learning (GSSL) is a practical solution to learn from a limited amount of labelled data together with a vast amount of unlabelled data. However, due to…
Efficient and Transferable Adversarial Examples from Bayesian Neural Networks
Martin Gubri, Maxime Cordy, Mike Papadakis +2
An established way to improve the transferability of black-box evasion attacks is to craft the adversarial examples on an ensemble-based surrogate to increase diversity. We argue t…
Adversarial Perturbation Intensity Achieving Chosen Intra-Technique Transferability Level for Logistic Regression
Martin Gubri
Machine Learning models have been shown to be vulnerable to adversarial examples, ie. the manipulation of data by a attacker to defeat a defender's classifier at test time. We pres…