5 citations · 6 across the 30 of their papers we have counts for
36 papers
Bridging Neural Tangent Kernel Analysis and Reinforcement Learning for Deep Function Approximation
William St-Arnaud, Margarida Carvalho, Golnoosh Farnadi
Despite their widespread success across machine learning, the learning dynamics and generalization properties of neural networks remain poorly understood. The neural tangent kernel…
Scaling Laws for Multimodal Data Mixtures
Aditi Khandelwal, Ayush K. Tarun, Yixuan Xu +5
Frontier AI systems are increasingly natively multimodal, jointly pretrained on multiple modalities, such as text, vision, and audio. However, the scientific understanding of the m…
Loose Lips: Understanding Privacy Leakage in Agentic Dialogue
Soumya Sharma, Ayana Hussain, Heber Hwang Arcolezi +4
As LLM agents are increasingly deployed in high-stakes enterprise settings, understanding how they handle sensitive information in agentic dialogue has become critical. Privacy lea…
LLM Persona Unlearning
Kemou Li, Zhuan Shi, Qizhou Wang +4
Pre-training equips large language models (LLMs) with a broad repertoire of behavioral patterns associated with roles, styles, values, and goals. Post-training teaches conditional…
Preemptive LLM Unlearning against Forbidden Capability Acquisition via Gradient Sealing
Kemou Li, Qizhou Wang, Yue Wang +6
Open-weight LLMs are released not only as fixed products but also as substrates for downstream fine-tuning. This openness, however, creates legal and ethical risks because users ma…
LLJ Cards: Best practices for the Use of LLMs as Judges
Khaoula Chehbouni, Melina Medjdoub, Florian Carichon +2
In recent years, large language models (LLMs) have emerged as a popular alternative for evaluation. Often referred to as LLMs as judges (LLJs), these systems have been widely adopt…
TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
Alireza Dehghanpour Farashah, Zhuan Shi, Negar Rostamzadeh +1
Text-to-image diffusion models can be misused to generate harmful content through adversarial or paraphrased prompts that bypass built-in safety mechanisms. Existing concept erasur…
IP Protection in the Era of Visual Generative AI: A Survey
Zhuan Shi, Shunchang Liu, Alireza Dehghanpour Farashah +8
The rapid evolution of visual generative AI has introduced a wide range of intellectual property risks, spanning the unauthorized learning, reproduction, extraction, misuse, and re…
Even More Deception: Objective Misalignment in Mixed-Motive LLM Multi-Agent Systems
Marylou Fauchard, Florian Carichon, Margarida Carvalho +1
Large Language Models (LLMs)-powered multi-agent systems are increasingly deployed in mixed-motive environments, where agents operate under asymmetric information and strategic dec…
Hidden in Thought: Transferable Chain-of-Thought Artifacts Induce Harmful Behavior
Ali khalil, Aly M. Kassem, Mohamed Abdelrazek +3
We investigate whether harmful chain-of-thought (CoT) traces from compromised language models can transfer unsafe behaviour and be distilled into reusable jailbreak attacks. Using…
IDP-Bench: Benchmarking ability of LLMs to protect personal information in interdependent privacy contexts
Ayana Hussain, Soumya Sharma, Golnoosh Farnadi +3
Large language models (LLMs) are becoming widely deployed as personal AI assistants with access to sensitive user data, making privacy a major challenge for their design and evalua…
Shared Latent Structures Enable Unified Backdoor Detection and Mitigation in LLMs
Omar Mahmoud, Aly M. Kassem, Thommen George Karimpanal +4
Backdoor attacks in large language models (LLMs) are often treated as isolated trigger-response failures, motivating defenses tailored to specific triggers or behaviors. We show th…
IDEAFix: Evaluation Framework for Creative Defixation Prompting in LLMs
F. Carichon, S. Sharma, M. Girard +2
Large language models (LLMs) are increasingly used for tasks involving creative problem solving and idea generation. However, there is a lack of consensus concerning their creative…
Leveraging Routing Dynamics in Mixture-of-Experts Models for Efficient Language Adaptation
Aditi Khandelwal, Marius Mosbach, Verna Dankers +2
Mixture-of-Experts (MoE) models are widely used to scale language models, yet their expert routing behavior and adaptation in a multilingual setting remain underexplored. In this w…
Mechanics of Bias and Reasoning: Interpreting the Impact of Chain-of-Thought Prompting on Gender Bias in LLMs
Edie Pearman, Sophia Osborne, Mira Kandlikar-Bloch +3
Large language models (LLMs) are increasingly deployed in socially sensitive settings despite substantial documentation that they encode gender biases. Chain-of-Thought (CoT) promp…
Neighbor-Aware Localized Concept Erasure in Text-to-Image Diffusion Models
Zhuan Shi, Alireza Dehghanpour Farashah, Rik de Vries +1
Concept erasure in text-to-image diffusion models seeks to remove undesired concepts while preserving overall generative capability. Localized erasure methods aim to restrict edits…
Delta-Crosscoder: Robust Crosscoder Model Diffing in Narrow Fine-Tuning Regimes
Aly Kassem, Thomas Jiralerspong, Negar Rostamzadeh +1
Model diffing methods aim to identify how fine-tuning changes a model's internal representations. Crosscoders approach this by learning shared dictionaries of interpretable latent…
Can LLMs Cook Jamaican Couscous? A Study of Cultural Novelty in Recipe Generation
F. Carichon, R. Rampa, G. Farnadi
Large Language Models (LLMs) are increasingly used to generate and shape cultural content, ranging from narrative writing to artistic production. While these models demonstrate imp…
Rethinking Hallucinations: Correctness, Consistency, and Prompt Multiplicity
Prakhar Ganesh, Reza Shokri, Golnoosh Farnadi
Large language models (LLMs) are known to "hallucinate" by generating false or misleading outputs. Hallucinations pose various harms, from erosion of trust to widespread misinforma…
Multilingual Amnesia: On the Transferability of Unlearning in Multilingual LLMs
Alireza Dehghanpour Farashah, Aditi Khandelwal, Marylou Fauchard +3
As multilingual large language models become more widely used, ensuring their safety and fairness across diverse linguistic contexts presents unique challenges. While existing rese…
Revisiting the Neural Tangent Kernel: the role of large width and depth
William St-Arnaud, Margarida Carvalho, Golnoosh Farnadi
Overparameterized fully-connected neural networks have been shown to behave like kernel models when trained with gradient descent, assuming standard scaling conditions on the width…
Data as a Lever: A Neighbouring Datasets Perspective on Predictive Multiplicity
Prakhar Ganesh, Hsiang Hsu, Golnoosh Farnadi
Multiplicity, the existence of equally good yet competing models, has received growing attention in recent years. While prior work has emphasized modelling choices, the critical ro…
Designing Ambiguity Sets for Distributionally Robust Optimization Using Structural Causal Optimal Transport
Ahmad-Reza Ehyaei, Golnoosh Farnadi, Samira Samadi
Distributionally robust optimization tackles out-of-sample issues like overfitting and distribution shifts by adopting an adversarial approach over a range of possible data distrib…
Wasserstein Distributionally Robust Optimization Through the Lens of Structural Causal Models and Individual Fairness
Ahmad-Reza Ehyaei, Golnoosh Farnadi, Samira Samadi
In recent years, Wasserstein Distributionally Robust Optimization (DRO) has garnered substantial interest for its efficacy in data-driven decision-making under distributional uncer…