https://github.com/aise-tudelft/llm4code-extraction

https://github.com/aise-tudelft/llm4code-extraction

Science Score: 26.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
  • codemeta.json file
    Found codemeta.json file
  • .zenodo.json file
    Found .zenodo.json file
  • DOI references
  • Academic publication links
  • Academic email domains
  • Institutional organization owner
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (7.6%) to scientific vocabulary
Last synced: 10 months ago · JSON representation

Repository

Basic Info
  • Host: GitHub
  • Owner: AISE-TUDelft
  • License: apache-2.0
  • Language: Python
  • Default Branch: main
  • Size: 52 MB
Statistics
  • Stars: 4
  • Watchers: 0
  • Forks: 1
  • Open Issues: 0
  • Releases: 0
Created almost 3 years ago · Last pushed about 1 year ago
Metadata Files
Readme License

README.md

Traces of Memorisation in Large Language Models for Code

Replication package for the paper titled "Traces of Memorisation in Large Language Models for Code"

For questions about the content of this repo, please use the issues board. If you have any questions about the paper, please email the first author.

Requirements

The requirements can be installed using Conda by running:

conda create --name LLMExtract --file requirements.txt

Make sure that you have added conda-forge to your conda channels:

conda config --append channels conda-forge

The code is intended to run on an Nvidia A40 with 48GB Vram, 32GB of RAM and 8 CPU cores. Smaller models can be run with lesser hardware as well, but be sure to change the batchsizes in the scripts.

Generation

To run an experiment simply run the following commands. The $model_name should be set to the path of a model on HuggingFace, such as Salesforce/codegen2-1B.

For text models:

python3 generatetext.py $modelname

For code models:

python3 generate.py $model_name

Ethical use

Use the code and concepts shared here ethically. The authors have shared this code to improve the security and safety of LLMs. Do not use this code for malicious purposes. When disclosing data leakage do not needlessly use put the privacy of people at risk.

Owner

  • Name: AISE-TUDelft
  • Login: AISE-TUDelft
  • Kind: organization

GitHub Events

Total
  • Push event: 1
Last Year
  • Push event: 1

Issues and Pull Requests

Last synced: over 1 year ago

All Time
  • Total issues: 0
  • Total pull requests: 0
  • Average time to close issues: N/A
  • Average time to close pull requests: N/A
  • Total issue authors: 0
  • Total pull request authors: 0
  • Average comments per issue: 0
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
Past Year
  • Issues: 0
  • Pull requests: 0
  • Average time to close issues: N/A
  • Average time to close pull requests: N/A
  • Issue authors: 0
  • Pull request authors: 0
  • Average comments per issue: 0
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
Top Authors
Issue Authors
Pull Request Authors
Top Labels
Issue Labels
Pull Request Labels