Skip to content

Responsible and Safe AI practices. Studying and recommending solutions for Identifying memorization by Large Language Models

Notifications You must be signed in to change notification settings

bssoft82/rsai-llm-memorization

Repository files navigation

Introduction

With the increasing power of LLMs, it is pretty easy for someone to infer, the PII data available through the information available on internet. This project compares various models output and also demonstrates the solutions and their effectiveness to overcome the challenge.

Steps to work with the project

Intialize the environment

python -m create venv .venv

. ./.venv/bin/activate

pip install -r requirements.txt

Change the openai api key

Execute the Program

python src/rsai.py

Viewing the results

CSV file

Open output/rsai_output.csv file and you can compare the output of models from this program

Log file

Open logs/rsai_logs.log to view the logs for each run

Folder Structure

The project has following folder structure:

  • config: contains configuration file for rsai program
  • data: contains input data for rsai program
  • logs: contains log file generated by rsai program
  • output: contains output file generated by rsai program
  • src: contains source code for rsai program

About

Responsible and Safe AI practices. Studying and recommending solutions for Identifying memorization by Large Language Models

Topics

Resources

Stars

Watchers

Forks

Packages

No packages published