From 58eccffe9d473134044ef937e04e837619d1a0ce Mon Sep 17 00:00:00 2001 From: Jamie Coombes Date: Sat, 27 Apr 2024 16:27:05 +0200 Subject: [PATCH 1/2] Create CODE_OF_CONDUCT.md --- CODE_OF_CONDUCT.md | 128 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 CODE_OF_CONDUCT.md diff --git a/CODE_OF_CONDUCT.md b/CODE_OF_CONDUCT.md new file mode 100644 index 0000000..d16b6ae --- /dev/null +++ b/CODE_OF_CONDUCT.md @@ -0,0 +1,128 @@ +# Contributor Covenant Code of Conduct + +## Our Pledge + +We as members, contributors, and leaders pledge to make participation in our +community a harassment-free experience for everyone, regardless of age, body +size, visible or invisible disability, ethnicity, sex characteristics, gender +identity and expression, level of experience, education, socio-economic status, +nationality, personal appearance, race, religion, or sexual identity +and orientation. + +We pledge to act and interact in ways that contribute to an open, welcoming, +diverse, inclusive, and healthy community. + +## Our Standards + +Examples of behavior that contributes to a positive environment for our +community include: + +* Demonstrating empathy and kindness toward other people +* Being respectful of differing opinions, viewpoints, and experiences +* Giving and gracefully accepting constructive feedback +* Accepting responsibility and apologizing to those affected by our mistakes, + and learning from the experience +* Focusing on what is best not just for us as individuals, but for the + overall community + +Examples of unacceptable behavior include: + +* The use of sexualized language or imagery, and sexual attention or + advances of any kind +* Trolling, insulting or derogatory comments, and personal or political attacks +* Public or private harassment +* Publishing others' private information, such as a physical or email + address, without their explicit permission +* Other conduct which could reasonably be considered inappropriate in a + professional setting + +## Enforcement Responsibilities + +Community leaders are responsible for clarifying and enforcing our standards of +acceptable behavior and will take appropriate and fair corrective action in +response to any behavior that they deem inappropriate, threatening, offensive, +or harmful. + +Community leaders have the right and responsibility to remove, edit, or reject +comments, commits, code, wiki edits, issues, and other contributions that are +not aligned to this Code of Conduct, and will communicate reasons for moderation +decisions when appropriate. + +## Scope + +This Code of Conduct applies within all community spaces, and also applies when +an individual is officially representing the community in public spaces. +Examples of representing our community include using an official e-mail address, +posting via an official social media account, or acting as an appointed +representative at an online or offline event. + +## Enforcement + +Instances of abusive, harassing, or otherwise unacceptable behavior may be +reported to the community leaders responsible for enforcement at +obvslib@protonmail.com. +All complaints will be reviewed and investigated promptly and fairly. + +All community leaders are obligated to respect the privacy and security of the +reporter of any incident. + +## Enforcement Guidelines + +Community leaders will follow these Community Impact Guidelines in determining +the consequences for any action they deem in violation of this Code of Conduct: + +### 1. Correction + +**Community Impact**: Use of inappropriate language or other behavior deemed +unprofessional or unwelcome in the community. + +**Consequence**: A private, written warning from community leaders, providing +clarity around the nature of the violation and an explanation of why the +behavior was inappropriate. A public apology may be requested. + +### 2. Warning + +**Community Impact**: A violation through a single incident or series +of actions. + +**Consequence**: A warning with consequences for continued behavior. No +interaction with the people involved, including unsolicited interaction with +those enforcing the Code of Conduct, for a specified period of time. This +includes avoiding interactions in community spaces as well as external channels +like social media. Violating these terms may lead to a temporary or +permanent ban. + +### 3. Temporary Ban + +**Community Impact**: A serious violation of community standards, including +sustained inappropriate behavior. + +**Consequence**: A temporary ban from any sort of interaction or public +communication with the community for a specified period of time. No public or +private interaction with the people involved, including unsolicited interaction +with those enforcing the Code of Conduct, is allowed during this period. +Violating these terms may lead to a permanent ban. + +### 4. Permanent Ban + +**Community Impact**: Demonstrating a pattern of violation of community +standards, including sustained inappropriate behavior, harassment of an +individual, or aggression toward or disparagement of classes of individuals. + +**Consequence**: A permanent ban from any sort of public interaction within +the community. + +## Attribution + +This Code of Conduct is adapted from the [Contributor Covenant][homepage], +version 2.0, available at +https://www.contributor-covenant.org/version/2/0/code_of_conduct.html. + +Community Impact Guidelines were inspired by [Mozilla's code of conduct +enforcement ladder](https://github.com/mozilla/diversity). + +[homepage]: https://www.contributor-covenant.org + +For answers to common questions about this code of conduct, see the FAQ at +https://www.contributor-covenant.org/faq. Translations are available at +https://www.contributor-covenant.org/translations. From bc06ba3e869a49bbbe8895f7bc9de646f7e49d41 Mon Sep 17 00:00:00 2001 From: Jamie Coombes Date: Sat, 27 Apr 2024 16:24:26 +0100 Subject: [PATCH 2/2] README and CONTRIBUTING guides --- CONTRIBUTING.md | 88 ++++++++++++++++++++++++++++++++++++++ README.md | 111 ++++++++++++++++++++++++++++++++++++++++-------- 2 files changed, 182 insertions(+), 17 deletions(-) create mode 100644 CONTRIBUTING.md diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md new file mode 100644 index 0000000..5e95e81 --- /dev/null +++ b/CONTRIBUTING.md @@ -0,0 +1,88 @@ +# Contributing to `obvs` + +Welcome to the `obvs` library! We appreciate your interest in contributing to the project. This document provides guidelines and information to help you contribute effectively. + +## Table of Contents + +- [Welcome Note](#welcome-note) +- [Links to Resources](#links-to-resources) + - [Templates](#templates) + - [Test Location](#test-location) + - [Submit Changes](#submit-changes) + - [Development Environment Details](#development-environment-details) +- [How To Contribute](#how-to-contribute) + - [How to Make a Bug Report](#how-to-make-a-bug-report) + - [How to Fix a Bug](#how-to-fix-a-bug) + - [How to Suggest Enhancements](#how-to-suggest-enhancements) +- [Coding Conventions and Style Guide](#coding-conventions-and-style-guide) +- [Code of Conduct](#code-of-conduct) +- [Acknowledgements](#acknowledgements) +- [Project Owner and Contributors Information](#project-owner-and-contributors-information) +- [Where to Get Help](#where-to-get-help) + +## Welcome Note + +Thank you for considering contributing to the `obvs` library! Your contributions, whether they are bug reports, feature suggestions, or code improvements, are greatly appreciated. By contributing, you help make `obvs` a better tool for analyzing and interpreting language models. + +## Links to Resources + +### Templates + +- [Bug Report Template](link-to-bug-report-template) +- [Feature Request Template](link-to-feature-request-template) +- [Pull Request Template](link-to-pull-request-template) + +### Test Location + +Please place your tests in the `tests/` directory and ensure they follow the naming convention `test_*.py`. + +### Submit Changes + +To submit your changes, please follow the steps outlined in the [How To Contribute](#how-to-contribute) section. + +### Development Environment Details + +To set up the development environment for `obvs`, please refer to the [Development Setup](https://github.com/obvslib/obvs?tab=readme-ov-file#development-setup) section in the README. + +## How To Contribute + +### How to Make a Bug Report + +If you encounter a bug while using `obvs`, please [open an issue](https://github.com/obvslib/obvs/issues) on the GitHub repository. Provide a clear and descriptive title, along with a detailed description of the problem, steps to reproduce it, and any relevant code snippets or error messages. + +### How to Fix a Bug + +If you would like to fix a bug, follow these steps: + +1. Check the [issue tracker](https://github.com/obvslib/obvs/issues) to see if the bug has already been reported. If not, [open a new issue](https://github.com/obvslib/obvs/issues/new) describing the bug. +2. Fork the repository and create a new branch for your bug fix. +3. Make the necessary changes to fix the bug. +4. Write tests to verify the bug fix and ensure it doesn't introduce new issues. +5. Commit your changes and push them to your forked repository. +6. [Submit a pull request](https://github.com/obvslib/obvs/compare) to the main repository, referencing the related issue. + +### How to Suggest Enhancements + +If you have an idea for an enhancement or a new feature for `obvs`, please [open an issue](https://github.com/obvslib/obvs/issues/new) on the GitHub repository. Provide a clear and descriptive title, along with a detailed description of the enhancement, its benefits, and any relevant examples or use cases. + +## Coding Conventions and Style Guide + +When contributing code to `obvs`, please adhere to the following coding conventions and style guide: + +- Use meaningful variable and function names. +- Follow the [PEP 8](https://www.python.org/dev/peps/pep-0008/) style guide for Python code. +- Write docstrings for functions and classes following the [NumPy docstring format](https://numpydoc.readthedocs.io/en/latest/format.html). +- Use type hints to improve code readability and maintainability. +- Keep lines of code under 80 characters in length. + +## Code of Conduct + +Please note that the `obvs` project follows a `CODE_OF_CONDUCT`. By participating in this project, you are expected to uphold this code. Please report any unacceptable behavior to the project maintainers. + +## Acknowledgements + +We would like to express our gratitude to all the contributors who have helped improve the `obvs` library. Your contributions are greatly appreciated! + +## Where to Get Help + +If you have any questions or need assistance with contributing to `obvs`, please [open an issue](https://github.com/obvslib/obvs/issues/new) on the GitHub repository. You can also reach out to the project maintainers or join our [github discussions](https://github.com/obvslib/obvs/discussions) for further help and discussion. diff --git a/README.md b/README.md index 990dcb8..ad877da 100644 --- a/README.md +++ b/README.md @@ -1,23 +1,100 @@ -# obvs +# obvs: A Python Library for Analyzing and Interpreting Language Models -[![CI](https://github.com/obvslib/obvs/actions/workflows/main.yaml/badge.svg)](https://github.com/obvslib/obvs/actions/workflows/main.yaml) +The `obvs` library is a powerful Python package that provides a comprehensive set of tools and utilities for analyzing and interpreting language models using the patchscope framework. It offers a range of functionalities to probe and understand the internal representations and behaviors of language models at different layers and positions. -Making Transformers Obvious +With `obvs`, you can easily investigate how language models process and generate text, gain insights into their inner workings, and visualize the results using various techniques such as heatmaps and plots. Whether you are a researcher, data scientist, or language model enthusiast, `obvs` empowers you to conduct interpretability experiments and reproduce standard results with ease. -## Project cheatsheet +## Installation -- **pre-commit:** `pre-commit run --all-files` -- **pytest:** `pytest` or `pytest -s` -- **coverage:** `coverage run -m pytest` or `coverage html` -- **poetry sync:** `poetry install --no-root --sync` -- **updating requirements:** see [docs/updating_requirements.md](docs/updating_requirements.md) +To install the `obvs` library, you can use pip: -## Initial project setup +``` +pip install obvs +``` -1. See [docs/getting_started.md](docs/getting_started.md) or [docs/quickstart.md](docs/quickstart.md) - for how to get up & running. -2. Check [docs/project_specific_setup.md](docs/project_specific_setup.md) for project specific setup. -3. See [docs/using_poetry.md](docs/using_poetry.md) for how to update Python requirements using - [Poetry](https://python-poetry.org/). -4. See [docs/detect_secrets.md](docs/detect_secrets.md) for more on creating a `.secrets.baseline` - file using [detect-secrets](https://github.com/Yelp/detect-secrets). +Make sure you have Python >=3.10 installed on your system. + +## Entity Resolution Example. + +Here's a simple example demonstrating how to use the `TokenIdentity` lens from the `obvs` library: + +```python +from obvs.lenses import TokenIdentity + +token_identity = TokenIdentity(source_prompt="The quick brown fox", model_name="gpt2", source_phrase="quick brown") +token_identity.run().compute_surprisal("fox").visualize() +``` + +This code snippet creates an instance of the `TokenIdentity` lens, specifying the source prompt, model name, and source phrase. It then runs the lens analysis, computes the surprisal for the word "fox", and visualizes the results. + +## Activation Patching Example. + +"The Collosseum is in the city of Paris" activation patching example. + +```python +from obvs.patchscope import Patchscope, SourceContext, TargetContext + +MODEL="gpt2" +source=SourceContext( + model_name=MODEL, + prompt="The Eiffel Tower is in the city of", + layer=10, + position=9, +) + +target=TargetContext( + model_name=MODEL, + prompt="The Colosseum is in the city of", + layer=10, + position=9, + max_new_tokens=1 +) + +patchscope=Patchscope(source, target) +patchscope.run() + +print(patchscope.full_output()) +``` + +For more examples and usage, please refer to the [documentation](https://obvs.rtfd.io/) and [PyPI](https://pypi.org/project/obvs/). + +## Development setup + +To set up the development environment for `obvs`, follow these steps: + +1. Clone the repository: + + ``` + git clone https://github.com/yourname/obvs.git + ``` + +2. Install the development dependencies: + + ``` + poetry install --no-root --sync + ``` + +3. Run the test suite: + ``` + pytest tests/ + ``` + +Make sure you have Python 3.10 or above and the required dependencies installed. + +## Release History + +- 0.1.1 + - Initial release of the `obvs` library + - Includes `patchscope`, `patchscope_base`, `lenses`, `logging`, and `metrics` modules + - Provides a collection of scripts for reproducing standard results + +## Meta + +For assistance, reach out to Jamie Coombes – www.linkedin.com/in/ +jamiecoombes – obvslib@protonmail.com + +Distributed under the MIT license. See `LICENSE` for more information. + +## Contributing + +We welcome contributions to the `obvs` library! See `CONTRIBUTING` for more information.