Skip to content
View jorispos's full-sized avatar
8 followers · 3 following

Achievements

Achievement: Pull Sharkx3Achievement: StarstruckAchievement: QuickdrawAchievement: YOLOAchievement: Pair Extraordinairex2Achievement: Arctic Code Vault Contributor

Achievements

Achievement: Pull Sharkx3Achievement: StarstruckAchievement: QuickdrawAchievement: YOLOAchievement: Pair Extraordinairex2Achievement: Arctic Code Vault Contributor

Highlights

  • Pro

Block or report jorispos

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
jorispos/README.md

Joris Postmus

I'm an AI researcher and developer doing R&D at Waking Up. My work spans AI research, engineering, and product development.

Most of my current engineering work lives in private repositories. My recent public research explores conceptors as a way to make language-model steering more precise and compositional.

🔬 Research

Steering Large Language Models using Conceptors

Research with Steven Abreu, presented at the NeurIPS 2024 Workshop on Foundation Model Interventions.

Paper | Code | Project page

🌐 Website | 💼 LinkedIn | 🎓 Google Scholar

Pinned Loading

  1. ConceptorSteering ConceptorSteering Public

    Official code for Steering Large Language Models using Conceptors, presented at the NeurIPS 2024 MINT Workshop.

    Jupyter Notebook 16 4

  2. Neural-Time-Series-Forecasting Neural-Time-Series-Forecasting Public

    Recursive multi-step forecasting with MLPs, polynomial detrending, and seasonal decomposition.

    Python 1

  3. NLP-Emotion-Classification NLP-Emotion-Classification Public

    TF-IDF text representations and Complement Naive Bayes for seven-class emotion recognition.

    Python 1

  4. Deep-RL-Lunar-Lander Deep-RL-Lunar-Lander Public

    Deep Q-learning with experience replay, benchmarked against SARSA and linear Q-function approximation.

    Python 1

  5. JPCode JPCode Public

    A visual text-encoding experiment I built in Java at sixteen, inspired by sci-fi robots communicating through images.

    Java