{"id":24725455,"url":"https://github.com/aise-tudelft/llm4code-memtune","last_synced_at":"2026-04-29T11:03:38.007Z","repository":{"id":272894774,"uuid":"917784511","full_name":"AISE-TUDelft/LLM4Code-memtune","owner":"AISE-TUDelft","description":"Replication package for the paper: \"How Much Do Code Language Models Remember? An Investigation on Data Extraction Attacks before and after Fine-tuning\"","archived":false,"fork":false,"pushed_at":"2025-09-11T12:56:54.000Z","size":14044,"stargazers_count":4,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-01-27T09:44:45.810Z","etag":null,"topics":["fine-tuning","llm4code","memorization","starcoder2"],"latest_commit_sha":null,"homepage":"","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/AISE-TUDelft.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2025-01-16T16:34:08.000Z","updated_at":"2025-10-09T11:21:53.000Z","dependencies_parsed_at":"2025-01-17T09:36:03.450Z","dependency_job_id":"f0d49c99-4bff-40f7-9a14-a28abcd66ac8","html_url":"https://github.com/AISE-TUDelft/LLM4Code-memtune","commit_stats":null,"previous_names":["aise-tudelft/llm4code-memtune"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/AISE-TUDelft/LLM4Code-memtune","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AISE-TUDelft%2FLLM4Code-memtune","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AISE-TUDelft%2FLLM4Code-memtune/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AISE-TUDelft%2FLLM4Code-memtune/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AISE-TUDelft%2FLLM4Code-memtune/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/AISE-TUDelft","download_url":"https://codeload.github.com/AISE-TUDelft/LLM4Code-memtune/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AISE-TUDelft%2FLLM4Code-memtune/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":32422546,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-29T06:29:02.080Z","status":"ssl_error","status_checked_at":"2026-04-29T06:29:00.631Z","response_time":110,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.6:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["fine-tuning","llm4code","memorization","starcoder2"],"created_at":"2025-01-27T13:18:35.915Z","updated_at":"2026-04-29T11:03:37.985Z","avatar_url":"https://github.com/AISE-TUDelft.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# LLM4Code-memtune\n\nReplication package for the paper: \"**How Much Do Code Language Models Remember? An Investigation on Data Extraction Attacks before and after Fine-tuning**\"\n\nFor questions:\n- Repository content: Please use the issues board\n- Paper inquiries: Contact the first author via email (info DOT fabiosalern AT gmail DOT COM)\n\n## Repository Structure\n\n```\nLLM4Code-memtune/\n├── data/               # Dataset filtering and sample creation tools\n├── training/          # StarCoder2 fine-tuning scripts and training stats\n└── evaluation/        # Data extraction experiment code and results\n```\n\n## Requirements\n\n### Hardware Requirements\n- GPU: Nvidia A100 (80GB VRAM)\n- RAM: 32GB\n- CPU: 16 cores\n\nGPU requirements by model:\n- StarCoder2-3B: 2 GPUs\n- StarCoder2-7B: 4 GPUs\n- StarCoder2-15B: 6 GPUs\n\nNote: Data extraction experiments can run on a single GPU.\n\n### Software Requirements\n- Python 3.8\n- Additional dependencies:\n  ```bash\n  pip install -r requirements.txt\n  ```\n\n## Directories\n\nFor detailed documentation of each directory, please refer to their respective README files.\n\n### Data\nContains scripts and tools for dataset filtering and sample creation, organized into two main directories.\n\n- The fine-tuning dataset can be retrieved at this link: [AISE-TUDelft/memtune-tuning_data](https://huggingface.co/datasets/AISE-TUDelft/memtune-tuning_data)\n\n### Training\nContains:\n- Fine-tuning scripts for StarCoder2\n- Training statistics and metrics\n\n- The fine-tuned models can be retrieved at this link: [AISE-TUDelft/LLM4Code-memtune](https://huggingface.co/collections/AISE-TUDelft/llm4code-memtune-678a2838766dd16037a8bfe0)\n\n### Evaluation\nContains code, data, and results for data extraction experiments.\n\n- The generated data extraction benchmarks are available at [AISE-TUDelft/memtune-data_attack](https://huggingface.co/datasets/AISE-TUDelft/memtune-data_attack)\n\n## Ethical use\nPlease use the code and concepts shared here responsibly and ethically. The authors have provided this code to enhance the security and safety of large language models (LLMs). Avoid using this code for any malicious purposes. When disclosing data leakage, take care not to compromise individuals' privacy unnecessarily.\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Faise-tudelft%2Fllm4code-memtune","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Faise-tudelft%2Fllm4code-memtune","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Faise-tudelft%2Fllm4code-memtune/lists"}