2 citations · 3 across the 7 of their papers we have counts for
7 papers
FLIPS: Instance-Fingerprinting for LLMs via Pseudo-random Sequences
Gurvan Richardeau, Gohar Dashyan, Erwan Le Merrer +1
Literature reveals that a Large Language Model's (LLM) behavior is not only conditioned by its original weights but also its instance-level parameters, such as instructional prompt…
Robust ML Auditing using Prior Knowledge
Jade Garcia Bourrée, Augustin Godinot, Martijn De Vos +5
Among the many technical challenges to enforcing AI regulations, one crucial yet underexplored problem is the risk of audit manipulation. This manipulation occurs when a platform d…
P2NIA: Privacy-Preserving Non-Iterative Auditing
Jade Garcia Bourrée, Hadrien Lautraite, Sébastien Gambs +3
The emergence of AI legislation has increased the need to assess the ethical compliance of high-risk AI systems. Traditional auditing methods rely on platforms' application program…
Queries, Representation & Detection: The Next 100 Model Fingerprinting Schemes
Augustin Godinot, Erwan Le Merrer, Camilla Penzo +2
The deployment of machine learning models in operational contexts represents a significant investment for any organisation. Consequently, the risk of these models being misappropri…
The 20 questions game to distinguish large language models
Gurvan Richardeau, Erwan Le Merrer, Camilla Penzo +1
In a parallel with the 20 questions game, we present a method to determine whether two large language models (LLMs), placed in a black-box context, are the same or not. The goal is…
LLMs Prompted for Graphs: Hallucinations and Generative Capabilities
Gurvan Richardeau, Samy Chali, Erwan Le Merrer +2
Large Language Models (LLMs) are nowadays prompted for a wide variety of tasks. In this article, we investigate their ability in reciting and generating graphs. We first study the…
Under manipulations, are some AI models harder to audit?
Augustin Godinot, Gilles Tredan, Erwan Le Merrer +2
Auditors need robust methods to assess the compliance of web platforms with the law. However, since they hardly ever have access to the algorithm, implementation, or training data…