Skip to content
@EleutherAI

EleutherAI

Pinned Loading

  1. gpt-neox Public

    An implementation of model parallel autoregressive transformers on GPUs, based on the Megatron and DeepSpeed libraries

    Python 7.1k 1k

  2. lm-evaluation-harness Public

    A framework for few-shot evaluation of language models.

    Python 8.5k 2.3k

  3. minetest Public

    Forked from luanti-org/luanti

    Minetest is an open source voxel game engine with easy modding and game creation

    C++ 64 11

  4. pythia Public

    The hub for EleutherAI's work on interpretability and learning dynamics

    Jupyter Notebook 2.4k 181

Repositories

Showing 10 of 160 repositories
  • DeeperSpeed Public Forked from deepspeedai/DeepSpeed

    DeepSpeed is a deep learning optimization library that makes distributed training easy, efficient, and effective.

    Python 164 Apache-2.0 4,491 0 2 Updated Mar 31, 2025
  • open-r1 Public Forked from huggingface/open-r1

    Fully open reproduction of DeepSeek-R1

    Python 0 Apache-2.0 2,146 0 0 Updated Mar 30, 2025
  • lm-evaluation-harness Public

    A framework for few-shot evaluation of language models.

    Python 8,457 MIT 2,259 362 (19 issues need help) 104 Updated Mar 30, 2025
  • tyche Public

    Precisely estimating the volume of basins in neural net parameter space corresponding to interpretable behaviors

    Jupyter Notebook 6 Apache-2.0 0 0 0 Updated Mar 29, 2025
  • sparsify Public

    Sparsify transformers with SAEs and transcoders

    Python 498 MIT 66 4 (1 issue needs help) 5 Updated Mar 28, 2025
  • gpt-neox Public

    An implementation of model parallel autoregressive transformers on GPUs, based on the Megatron and DeepSpeed libraries

    Python 7,148 Apache-2.0 1,047 64 (3 issues need help) 24 Updated Mar 24, 2025
  • elk Public

    Keeping language models honest by directly eliciting knowledge encoded in their activations.

    Python 197 MIT 33 15 (1 issue needs help) 10 Updated Mar 24, 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 165 Apache-2.0 23 2 3 Updated Mar 24, 2025
  • rtopk Public

    https://github.com/xiexi51/RTopK PyTorch wrapper

    Cuda 1 MIT 0 0 0 Updated Mar 22, 2025
  • POSER Public Forked from sevdeawesome/POSER

    Poser: Unmasking Alignment Faking LLMs by Manipulating Their Internals

    Python 0 1 0 0 Updated Mar 21, 2025
View all repositories