Method for using a large language model by a user, and associated learning method, learning and use method, and learning and use system

The method addresses the lack of fine-grained access control in large language models by creating confidential models with varying access levels and implementing user-specific decryption, ensuring data confidentiality and reducing unauthorized access risks.

FR3168995A1Pending Publication Date: 2026-05-29COMMISSARIAT A LENERGIE ATOMIQUE ET AUX ENERGIES ALTERNATIVES
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
FR · FR
Patent Type
Applications
Current Assignee / Owner
COMMISSARIAT A LENERGIE ATOMIQUE ET AUX ENERGIES ALTERNATIVES
Filing Date
2024-11-28
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

Existing methods for protecting the confidentiality of training data in large language models do not allow for effective fine-grained access control during the training, adaptation, and use phases, leaving sensitive data vulnerable to unauthorized access and data leak attacks.

Method used

A method and system for training and using a large language model that involves creating multiple confidential models, each associated with a specific level of confidentiality, and implementing fine-grained access control by evaluating user access rights to determine the appropriate model for interaction, with optional decryption based on user rights.

Benefits of technology

Ensures the confidentiality of training data by enabling differentiated access control, reducing the risk of unauthorized data access and protecting sensitive information throughout the model's lifecycle.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

Method of using a large language model by a user, and associated learning method, learning and use method and learning and use system. The present invention relates to a method of using a large language model by a user, the large language model being learned on training data of different levels of confidentiality and taking the form of a plurality of confidential models, each confidential model being associated with a level of confidentiality; the method of use comprising the following steps: - acquisition (310) of a request for use from the user; - evaluation (320) of the user's access rights, the access rights determining a level of confidentiality of the training data accessible by the user;- retrieval (330) of a confidential model corresponding to the level of confidentiality determined by the user's access rights; - implementation (340) of user interactions with the retrieved confidential model. Figure for the abbreviation: Figure 7;
Need to check novelty before this filing date? Find Prior Art