{"id":27400839,"url":"https://github.com/butanium/self-obf-interp","last_synced_at":"2026-02-18T07:32:31.206Z","repository":{"id":285663835,"uuid":"956503084","full_name":"Butanium/self-obf-interp","owner":"Butanium","description":null,"archived":false,"fork":false,"pushed_at":"2025-04-02T01:24:39.000Z","size":27552,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-10-20T09:47:03.706Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"HTML","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/Butanium.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"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,"zenodo":null}},"created_at":"2025-03-28T11:10:19.000Z","updated_at":"2025-04-02T01:24:42.000Z","dependencies_parsed_at":null,"dependency_job_id":"ba0d1bfe-2866-4980-88dd-a9020586f46b","html_url":"https://github.com/Butanium/self-obf-interp","commit_stats":null,"previous_names":["butanium/self-obf-interp"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/Butanium/self-obf-interp","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Butanium%2Fself-obf-interp","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Butanium%2Fself-obf-interp/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Butanium%2Fself-obf-interp/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Butanium%2Fself-obf-interp/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Butanium","download_url":"https://codeload.github.com/Butanium/self-obf-interp/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Butanium%2Fself-obf-interp/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":29572417,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-02-18T06:19:27.422Z","status":"ssl_error","status_checked_at":"2026-02-18T06:18:44.348Z","response_time":162,"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":[],"created_at":"2025-04-14T03:43:23.191Z","updated_at":"2026-02-18T07:32:31.201Z","avatar_url":"https://github.com/Butanium.png","language":"HTML","funding_links":[],"categories":[],"sub_categories":[],"readme":"All files in the root folder are mine. Notebook_results contains my plots. Data contains my custom gemma chat templates. The rest is from a version of https://github.com/mgm52/self-obfuscation\n\nTo run the code you'll need access to private models from Max, so reach out if you want to run it :)\n\n## Setup\n```bash\ngit config --global credential.helper store\npip install huggingface_hub uv\nhuggingface-cli login --token \u003cYOUR_ACCESS_TOKEN_WITH_GEMMA_AND_MAX_REPO\u003e --add-to-git-credential\napt update \u0026\u0026 apt install git-lfs -y\ngit lfs install\ngit-lfs clone https://huggingface.co/self-obfuscation-model-viewers/self-obf-files\nuv sync\ncd self-obf-files\ngit switch models-only\ncd ..\nuv run python -c \"from transformers import AutoModelForCausalLM; AutoModelForCausalLM.from_pretrained('google/gemma-2-9b-it')\"  # optional but will download gemma-2-9b-it if it's not in your cache already\n```\n\n## File Descriptions\n\n- `probe_experiments_notebook.ipynb`: Python notebook for running experiments to analyze self-obfuscation in language models using probing techniques. Includes layer stitching experiments, representation analysis, and visualization.\n- `data_utils.py`: Utilities for loading and processing test data for self-obfuscation experiments, including functions to create variations of examples with different adversarial words.\n- `utils.py`: General utility functions for self-obfuscation experiments, including functions for patching model activations, extracting token positions, and layer stitching.\n- `visualization.py`: Extensive visualization utilities for plotting experimental results, including layer sweep visualization, representation norm plots, KL divergence visualization, and token probability visualizations.\n- `probes.py`: Functions for loading and evaluating probes to detect specific concepts in model representations.\n- `models.py`: Model loading and patching utilities, including functions to load checkpoints, stitch model layers, and extract model representations.","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbutanium%2Fself-obf-interp","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fbutanium%2Fself-obf-interp","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbutanium%2Fself-obf-interp/lists"}