Martin Gubri's picture

Martin Gubri

mgubri

·

https://gubri.eu/

AI & ML interests

Adversarial Examples, Machine Learning, Information Security, Statistics

Organizations

upvoted a collection 4 months ago

Gaperon

Our French-English LLM suite (including Base and SFT models. All checkpoints are also included. • 16 items • Updated 2 days ago • 17

upvoted a paper 4 months ago

Privacy Collapse: Benign Fine-Tuning Can Break Contextual Privacy in Language Models

Paper • 2601.15220 • Published Jan 21 • 9

upvoted 8 papers 7 months ago

Is Multilingual LLM Watermarking Truly Multilingual? A Simple Back-Translation Solution

Paper • 2510.18019 • Published Oct 20, 2025 • 18

C-SEO Bench: Does Conversational SEO Work?

Paper • 2506.11097 • Published Jun 6, 2025 • 2

Scaling Up Membership Inference: When and How Attacks Succeed on Large Language Models

Paper • 2411.00154 • Published Oct 31, 2024 • 1

TRAP: Targeted Random Adversarial Prompt Honeypot for Black-Box Identification

Paper • 2402.12991 • Published Feb 20, 2024 • 1

Calibrating Large Language Models Using Their Generations Only

Paper • 2403.05973 • Published Mar 9, 2024 • 1

ProPILE: Probing Privacy Leakage in Large Language Models

Paper • 2307.01881 • Published Jul 4, 2023 • 2

Dr.LLM: Dynamic Layer Routing in LLMs

Paper • 2510.12773 • Published Oct 14, 2025 • 32

DISCO: Diversifying Sample Condensation for Efficient Model Evaluation

Paper • 2510.07959 • Published Oct 9, 2025 • 15

upvoted a paper 9 months ago

Leaky Thoughts: Large Reasoning Models Are Not Private Thinkers

Paper • 2506.15674 • Published Jun 18, 2025 • 2

upvoted a collection about 1 year ago

MIA-Pile

Samples used for the NAACL 2025 Findings paper: "Scaling Up Membership Inference: When and How Attacks Succeed on Large Language Models." • 23 items • Updated Feb 3, 2025 • 1

upvoted 4 collections over 1 year ago

📈 Scaling MIA Data & Results

NAACL 2025 Findings "Scaling Up Membership Inference: When and How Attacks Succeed on Large Language Models" https://arxiv.org/abs/2411.00154 • 22 items • Updated Jun 5, 2025 • 2

📚 Papers

List of research articles of Parameter Lab • 9 items • Updated Jan 22 • 1

Model with Circuit Breakers

SoTA models with circuit breakers inserted. Top safety performance without losing capabilities. • 3 items • Updated Oct 25, 2024 • 5

🍑 Apricot Models

Fine-tuned models for black-box LLM calibration, trained for "Apricot: Calibrating Large Language Models Using Their Generations Only" (ACL 2024) • 9 items • Updated Nov 20, 2024 • 3