understanding-search / maze-transformerLinks
This repo is built to facilitate the training and analysis of autoregressive transformers on maze-solving tasks.
☆31Updated 10 months ago
Alternatives and similar repositories for maze-transformer
Users that are interested in maze-transformer are comparing it to the libraries listed below
Sorting:
- Interpreting how transformers simulate agents performing RL tasks☆87Updated last year
- Latent Program Network (from the "Searching Latent Program Spaces" paper)☆91Updated 4 months ago
- maze datasets for investigating OOD behavior of ML systems☆48Updated last month
- ☆37Updated 9 months ago
- ☆51Updated last year
- unofficial re-implementation of "Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasets"☆77Updated 3 years ago
- Official codebase for "Sampling For Learnability", published at NeurIPS 2024☆16Updated last month
- Scaling scaling laws with board games.☆49Updated 2 years ago
- Code for "Evidence of Learned Look-Ahead in a Chess-Playing Neural Network"☆24Updated last year
- An Open-Ended Agentic Simulator☆51Updated 11 months ago
- ☆26Updated 2 years ago
- see github.com/understanding-search/maze-transformer☆10Updated last year
- Official Implementation of `An Optimisation Framework for Unsupervised Environment Design` from RLC 2025☆12Updated this week
- Benchmarking RL for POMDPs in Pure JAX [Code for "Structured State Space Models for In-Context Reinforcement Learning" (NeurIPS 2023)]☆107Updated last year
- Bootstrapping ARC☆133Updated 7 months ago
- ☆66Updated last year
- Notebooks accompanying Anthropic's "Toy Models of Superposition" paper☆127Updated 2 years ago
- Mitigating Partial Observability in Sequential Decision Processes via the Lambda Discrepancy☆18Updated 8 months ago
- Code for the paper "Learning Temporal Distances: Contrastive Successor Features Can Provide a Metric Structure for Decision-Making"☆27Updated last year
- Emergent world representations: Exploring a sequence model trained on a synthetic task☆182Updated 2 years ago
- TPU pod commander is a package for managing and launching jobs on Google Cloud TPU pods.☆20Updated last year
- Minimal but scalable implementation of large language models in JAX☆35Updated last week
- A TinyStories LM with SAEs and transcoders☆12Updated 3 months ago
- Atari-style POMDPs☆11Updated this week
- A library for efficient patching and automatic circuit discovery.☆70Updated 2 months ago
- Sparse Autoencoder Training Library☆53Updated 2 months ago
- Tools for studying developmental interpretability in neural networks.☆99Updated 3 weeks ago
- ☆23Updated last year
- Simple JAX Graphics Library.☆36Updated 8 months ago
- Efficient World Models with Context-Aware Tokenization. ICML 2024☆105Updated 9 months ago