This was designed for interp researchers who want to do research on or with interp agents to give quality of life improvements and fix some of the annoying things you get from only using Claude cod…

https://attention-motifs.github.io

NVARC solution to ARC-AGI-2

A Spotify player in the terminal with full feature parity

Code for "Physics Steering: Causal Control of Cross-Domain Concepts in a Physics Foundation Model" paper.

Let's try to reproduce the Golden Gate Claude demo, but using open-source Llama 3.1 8B, and the Eiffel Tower.

Modular toolkit for Interactive Tutoring Systems. SRS++

Live-SWE-agent: live, runtime self-evolving software engineering agent

Quick illustration of how one can easily read books together with LLMs. It's great and I highly recommend it.

🎬 A curated list of movies every hacker & cyberpunk must watch.

PyTorch implementation of JiT https://arxiv.org/abs/2511.13720

Research projects

Modular Multi-Agent System for Scientific Research Assistance

Code for reproducing our paper "Are Sparse Autoencoders Useful? A Case Study in Sparse Probing"

studying (self-)supervised representations of Euclid galaxy imaging via SAEs

Code for Learning to Interpret Weight Differences in Language Models (Goel et al. 2025)

Official PyTorch Implementation of "Diffusion Transformers with Representation Autoencoders"

🌈 React for interactive command-line apps

The best ChatGPT that $100 can buy.

A suite of interpretability tasks to evaluate agents using Scribe for notebook access

Enhanced ChatGPT Clone: Features Agents, MCP, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message…

Optimize prompts, code, and more with AI-powered Reflective Text Evolution

Code for reproducing our paper "Not All Language Model Features Are Linear"