{"id":21918306,"url":"https://github.com/poppingtonic/transformer-visualization","last_synced_at":"2025-04-19T11:41:44.681Z","repository":{"id":63877295,"uuid":"562917628","full_name":"poppingtonic/transformer-visualization","owner":"poppingtonic","description":"Mechanistic Interpretability Tutorials, Results and research log as I learn from @neelnanda-io's wonderful Easy-Transformer","archived":false,"fork":false,"pushed_at":"2023-09-13T20:07:35.000Z","size":5398,"stargazers_count":3,"open_issues_count":2,"forks_count":2,"subscribers_count":4,"default_branch":"main","last_synced_at":"2023-09-14T11:04:39.460Z","etag":null,"topics":["gradio-interface","interpretability-jam","interpretable-ai","transformers","visualization"],"latest_commit_sha":null,"homepage":"","language":"Jupyter Notebook","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/poppingtonic.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}},"created_at":"2022-11-07T14:28:39.000Z","updated_at":"2023-09-09T05:36:17.000Z","dependencies_parsed_at":"2023-02-17T05:15:18.834Z","dependency_job_id":null,"html_url":"https://github.com/poppingtonic/transformer-visualization","commit_stats":null,"previous_names":[],"tags_count":0,"template":null,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/poppingtonic%2Ftransformer-visualization","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/poppingtonic%2Ftransformer-visualization/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/poppingtonic%2Ftransformer-visualization/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/poppingtonic%2Ftransformer-visualization/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/poppingtonic","download_url":"https://codeload.github.com/poppingtonic/transformer-visualization/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":226979191,"owners_count":17712573,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","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":["gradio-interface","interpretability-jam","interpretable-ai","transformers","visualization"],"created_at":"2024-11-28T19:50:39.812Z","updated_at":"2024-11-28T19:50:40.478Z","avatar_url":"https://github.com/poppingtonic.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Learning Mechanistic Interpretability on Transformers with EasyTransformer (now TransformerLens)\n\n\n_by Brian Muhia_\n\nFahamu, Inc\n\n\nThis repository houses the beginnings of a tutorial on mechanistic interpretability for Transformer language models.\n\n#### Pedagogy\nSo far, we have:\n1. Published a usable visualiser for tokens, fashioned from the `Hacky Interactive Lexoscope` by Neel Nanda.\n1. Written notes from rewriting `EasyTransformer_Demo.ipynb` by Neel, in order to learn the library and how to use it.\n\n#### Output\n1. Applied some tools and ideas in the demo towards observing induction heads in [`SOLU-8l-old`](https://transformer-circuits.pub/2022/solu/index.html), also trained by Neel.\n1. Generated IOI-style datasets:\n    - pkl_ioi_data.pkl is 100000 rows of IOI sentences from `ABBA` templates, most of which use multi-token terms.\n    - https://huggingface.co/datasets/fahamu/ioi\n        + mecha_ioi_26m.parquet is 26,010,000 rows of IOI sentences, mixing ABBA and BABA templates\n        + mecha_ioi_200k.parquet is 200,000 rows of IOI sentences, mixing ABBA and BABA templates\n\nAll inspired by the paper _Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small_, from Redwood Research. We are not affiliated with Redwood Research, and release this dataset to contribute to the collective research effort behind understanding how Transformer language models perform this task. \n\n###### With thanks and Acknowledgements:\n\n- Esben Kran, Sabrina Zaki - for hosting the Interpretability Jam, which accelerated this work.\n- Neel Nanda - for publishing TransformerLens and making public his research process. Wonderful gifts!\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fpoppingtonic%2Ftransformer-visualization","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fpoppingtonic%2Ftransformer-visualization","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fpoppingtonic%2Ftransformer-visualization/lists"}