Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 17537
    Prev
    Next
    INDEX
    Explanations

    mentions of body parts

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "ar" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Lumpur
    -0.87
    zinski
    -0.81
     Showdown
    -0.81
    rome
    -0.76
     Dew
    -0.76
    worth
    -0.73
    cade
    -0.70
     Mull
    -0.70
     Dull
    -0.67
     Rasmussen
    -0.66
    POSITIVE LOGITS
     ar
    3.69
     Ar
    1.60
    Ar
    1.57
     Archer
    1.29
     arch
    1.28
     AR
    1.16
     Ark
    1.16
     ank
    1.08
     arrow
    1.06
     bows
    1.06
    Activations Density 0.011%

    No Known Activations