© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 0-RES-JB
    11. 5071
    Prev
    Next
    INDEX
    Explanations

    repetitive patterns of words or phrases, specifically those having the prefix "rep" followed by a number

    oai_token-act-pair · gpt-3.5-turbo

    references to representatives, specifically in a political context

    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
     Holmes
    -0.74
     sealing
    -0.71
     tablets
    -0.69
     Golem
    -0.66
     head
    -0.65
     Parker
    -0.65
     TAMADRA
    -0.63
     sealed
    -0.63
     heading
    -0.63
     whales
    -0.62
    POSITIVE LOGITS
    rep
    4.56
    repre
    1.58
    Rep
    1.52
    REP
    1.44
     Rep
    1.27
    rev
    1.24
    reb
    1.24
     rep
    1.18
    re
    1.14
     Reps
    1.13
    Activations Density 0.009%

    No Known Activations