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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 14455
    Prev
    Next
    INDEX
    Explanations

    the activation of documents starting with the word "Read", possibly indicating the end of an article or a section

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    373
    +0.16
    0.7%
    651
    +0.13
    0.6%
    420
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    491
    +0.16
    0.01
    755
    +0.13
    0.01
    651
    +0.12
    0.01
    Negative Logits
    IDA
    -0.69
    naire
    -0.65
    ult
    -0.62
    xon
    -0.62
    antle
    -0.61
    -"
    -0.61
     Reborn
    -0.61
     Nept
    -0.61
    agall
    -0.60
     ILCS
    -0.60
    POSITIVE LOGITS
    htaking
    0.76
    hran
    0.66
    gments
    0.65
    iness
    0.65
    np
    0.64
    cloneembedreportprint
    0.62
     ï
    0.61
     Read
    0.61
    Write
    0.58
    andowski
    0.58
    Activations Density 0.017%

    No Known Activations