wattenberg / superpositionLinks
Code associated to papers on superposition (in ML interpretability)
☆31Updated 2 years ago
Alternatives and similar repositories for superposition
Users that are interested in superposition are comparing it to the libraries listed below
Sorting:
- The simplest, fastest repository for training/finetuning medium-sized GPTs.☆153Updated 2 months ago
- Notebooks accompanying Anthropic's "Toy Models of Superposition" paper☆129Updated 2 years ago
- ☆166Updated 2 years ago
- Understand and test language model architectures on synthetic tasks.☆222Updated last month
- Code Release for "Broken Neural Scaling Laws" (BNSL) paper☆59Updated last year
- nanoGPT-like codebase for LLM training☆102Updated 3 months ago
- ☆87Updated last year
- ☆27Updated 2 years ago
- Understanding how features learned by neural networks evolve throughout training☆36Updated 10 months ago
- Language models scale reliably with over-training and on downstream tasks☆98Updated last year
- Yet another random morning idea to be quickly tried and architecture shared if it works; to allow the transformer to pause for any amount…☆53Updated last year
- ☆53Updated last year
- ☆83Updated last year
- ☆45Updated last year
- Simple and efficient pytorch-native transformer training and inference (batched)☆78Updated last year
- unofficial re-implementation of "Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasets"☆79Updated 3 years ago
- ☆126Updated last year
- A MAD laboratory to improve AI architecture designs 🧪☆125Updated 8 months ago
- Code for NeurIPS 2024 Spotlight: "Scaling Laws and Compute-Optimal Training Beyond Fixed Training Durations"☆81Updated 10 months ago
- Code for the paper "The Impact of Positional Encoding on Length Generalization in Transformers", NeurIPS 2023☆136Updated last year
- Sparse and discrete interpretability tool for neural networks☆63Updated last year
- ☆37Updated last year
- [NeurIPS 2023] Learning Transformer Programs☆163Updated last year
- Erasing concepts from neural representations with provable guarantees☆232Updated 7 months ago
- Code to reproduce "Transformers Can Do Arithmetic with the Right Embeddings", McLeish et al (NeurIPS 2024)☆191Updated last year
- Unofficial but Efficient Implementation of "Mamba: Linear-Time Sequence Modeling with Selective State Spaces" in JAX☆87Updated last year
- Universal Neurons in GPT2 Language Models☆30Updated last year
- ☆56Updated 10 months ago
- Experiments for efforts to train a new and improved t5☆76Updated last year
- A set of Python scripts that makes your experience on TPU better☆54Updated last year