Skip to content

Teaching materials for use with the web-based Surprisal Toolkit, in tutorials on information theory and calculating surprisal from large language models.

License

Notifications You must be signed in to change notification settings

uds-lsv/surprisal-toolkit-teaching-materials

Repository files navigation

Surprisal Toolkit Teaching Materials

Here we share teaching materials from our paper "An Interactive Toolkit for Approachable NLP" by AriaRay Brown, Julius Steuer, Marius Mosbach, and Dietrich Klakow, for the TeachNLP Workshop at ACL 2024.

The slides and Jupyter notebooks were used to teach 3 tutorial sessions with the Surprisal Toolkit.

Teaching Materials

🍎 Presentation slides: Surprisal_toolkit_tutorial_slides.pdf

  • We used these for the workshop tutorial to present background on information theory and calculating surprisal from large language models.

🚀 Notebook tutorial with advanced coding exercises: tutorial_course_session_solution.ipynb

  • Suitable for more experienced coders.

💡 Notebook tutorial with coding exercises solution:

  • Contains coding solutions for tutorial_course_session.ipynb

Notebook tutorial with light coding exercises: tutorial_workshop.ipynb

  • Suitable for less experienced coders or less intensive self-learning events.

License

See the LICENSE file for license rights and limitations (MIT). We welcome the shared usage of our materials for educational purposes.

About

Teaching materials for use with the web-based Surprisal Toolkit, in tutorials on information theory and calculating surprisal from large language models.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published