NeurIPS 2024 LLM Privacy Challenge

In the AI era, Large Language Models (LLMs) drive many applications, yet their use of extensive data poses privacy risks. This competition seeks to pinpoint these risks across LLM stages and develop strategies for preserving privacy.

Objectives

The primary goal of this competition is to stimulate research and innovation in the field of LLM privacy, with a focus on: Identifying vulnerabilities that could lead to privacy breaches in LLMs. Developing and evaluating techniques to enhance the privacy of these models.

Important Dates

2024-07-01
23:59:59 UTC+8
Registration Open
New team registration is now closed.
2024-08-01
23:59:59 UTC+8
Registration Deadline
Registered teams that are not participating in CodaBench
can still participate the Red Team or Blue Team
2024-08-10
23:59:59 UTC+8
Release of Starting Kit and Datasets
2024-08-19
23:59:59 UTC+8
Release of Starting Kit for Red Team and Blue Team (Datasets are included)
2024-11-01
23:59:59 UTC+8
Submission Deadline
2024-11-15
23:59:59 UTC+8
Evaluating Submissions and Releasing Results