Skip to content

@EleutherAI EleutherAI

Change the repository type filter

All

    Repositories list

    166 repositories

    • lm-evaluation-harness

      Public
      A framework for few-shot evaluation of language models.
      transformerlanguage-modelevaluation-framework
      Python
      •
      MIT License
      •2.4k•8.9k•395•120•Updated May 6, 2025May 6, 2025

    • pythia

      Public
      The hub for EleutherAI's work on interpretability and learning dynamics
      Jupyter Notebook
      •
      Apache License 2.0
      •184•2.5k•26•3•Updated May 6, 2025May 6, 2025

    • website

      Public
      New website for EleutherAI based on Hugo static site generator
      HTML
      •5•4•0•3•Updated May 5, 2025May 5, 2025

    • elk

      Public
      Keeping language models honest by directly eliciting knowledge encoded in their activations.
      Python
      •
      MIT License
      •33•201•15•10•Updated May 5, 2025May 5, 2025

    • sparsify

      Public
      Sparsify transformers with SAEs and transcoders
      Python
      •
      MIT License
      •71•524•0•1•Updated May 5, 2025May 5, 2025

    • delphi

      Public
      Delphi was the home of a temple to Phoebus Apollo, which famously had the inscription, 'Know Thyself.' This library lets language models know themselves through automated interpretability.
      Python
      •
      Apache License 2.0
      •25•170•4•2•Updated May 5, 2025May 5, 2025

    • aria-amt

      Public
      Efficient and robust implementation of seq-to-seq automatic piano transcription.
      Python
      •
      Apache License 2.0
      •9•41•0•0•Updated May 5, 2025May 5, 2025

    • wmdp

      Public
      WMDP is a LLM proxy benchmark for hazardous knowledge in bio, cyber, and chemical security. We also release code for RMU, an unlearning method which reduces LLM performance on WMDP while retaining general capabilities.
      Jupyter Notebook
      •
      MIT License
      •32•0•0•0•Updated May 5, 2025May 5, 2025

    • gpt-neox

      Public
      An implementation of model parallel autoregressive transformers on GPUs, based on the Megatron and DeepSpeed libraries
      transformerslanguage-modelgpt-3deepspeed-library
      Python
      •
      Apache License 2.0
      •1.1k•7.2k•64•27•Updated May 3, 2025May 3, 2025

    • nanoGPT-mup

      Public
      The simplest, fastest repository for training/finetuning medium-sized GPTs.
      Python
      •
      MIT License
      •6.8k•107•0•0•Updated May 2, 2025May 2, 2025

    • truffaldino

      Public
      Investigating goal instability in RL
      Python
      •
      MIT License
      •0•0•0•0•Updated May 2, 2025May 2, 2025

    • attribute

      Public
      JavaScript
      •1•0•0•0•Updated May 1, 2025May 1, 2025

    • cookbook

      Public
      Deep learning for dummies. All the practical details and useful utilities that go into working with real models.
      Python
      •
      Apache License 2.0
      •40•789•8•0•Updated Apr 30, 2025Apr 30, 2025

    • tyche

      Public
      Precisely estimating the volume of basins in neural net parameter space corresponding to interpretable behaviors
      Jupyter Notebook
      •
      Apache License 2.0
      •0•8•0•1•Updated Apr 30, 2025Apr 30, 2025

    • open-r1

      Public
      Fully open reproduction of DeepSeek-R1
      Python
      •
      Apache License 2.0
      •2.2k•2•0•0•Updated Apr 30, 2025Apr 30, 2025

    • fmri

      Public
      Analogue of fMRI on artificial neural networks
      MIT License
      •0•2•0•0•Updated Apr 24, 2025Apr 24, 2025

    • POSER

      Public
      Poser: Unmasking Alignment Faking LLMs by Manipulating Their Internals
      Python
      •1•1•0•0•Updated Apr 24, 2025Apr 24, 2025

    • rllm

      Public
      Democratizing Reinforcement Learning for LLMs
      Jupyter Notebook
      •
      MIT License
      •295•0•0•0•Updated Apr 16, 2025Apr 16, 2025

    • Megatron-LM

      Public
      Ongoing research training transformer models at scale
      Python
      •
      Other
      •2.7k•0•0•0•Updated Apr 15, 2025Apr 15, 2025

    • rtopk

      Public
      https://github.com/xiexi51/RTopK PyTorch wrapper
      Cuda
      •
      MIT License
      •0•1•0•0•Updated Apr 5, 2025Apr 5, 2025

    • DeeperSpeed

      Public
      DeepSpeed is a deep learning optimization library that makes distributed training easy, efficient, and effective.
      Python
      •
      Apache License 2.0
      •4.4k•167•0•1•Updated Apr 1, 2025Apr 1, 2025

    • ccs

      Public
      Python
      •
      MIT License
      •6•6•1•4•Updated Mar 21, 2025Mar 21, 2025

    • mixture-of-depths

      Public
      MIT License
      •0•0•0•0•Updated Mar 17, 2025Mar 17, 2025

    • cupbearer

      Public
      A library for mechanistic anomaly detection
      Jupyter Notebook
      •
      MIT License
      •10•6•0•0•Updated Feb 26, 2025Feb 26, 2025

    • clearnets

      Public
      Python
      •
      MIT License
      •0•4•0•0•Updated Feb 18, 2025Feb 18, 2025

    • polyapprox

      Public
      Closed-form polynomial approximations to neural networks
      Python
      •
      MIT License
      •1•13•0•1•Updated Jan 31, 2025Jan 31, 2025

    • transformer-reasoning

      Public
      Experiments in transformer knowledge and reasoning
      Jupyter Notebook
      •
      MIT License
      •20•10•0•0•Updated Jan 30, 2025Jan 30, 2025

    • TransformerEngine

      Public
      A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit floating point (FP8) precision on Hopper and Ada GPUs, to provide better performance with lower memory utilization in both training and inference.
      Python
      •
      Apache License 2.0
      •416•0•0•0•Updated Jan 29, 2025Jan 29, 2025

    • sae_overlap

      Public
      Acompanying code for our research on SAE feature overlap when trained on different seeds.
      Jupyter Notebook
      •
      Apache License 2.0
      •1•3•0•0•Updated Jan 28, 2025Jan 28, 2025

    • mdl

      Public
      Minimum Description Length probing for neural network representations
      Python
      •
      MIT License
      •2•19•0•2•Updated Jan 28, 2025Jan 28, 2025