© 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 · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 10500
    Prev
    Next
    INDEX
    Explanations

    punctuation marks occurring in sequences of three

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    79
    +0.10
    0.4%
    509
    +0.09
    0.4%
    160
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    592
    +0.10
    0.01
    509
    +0.09
    0.01
    206
    +0.09
    0.01
    Negative Logits
    ãĥ´
    -0.72
    ãĥ¼ãĥĨ
    -0.61
    nen
    -0.60
     welcome
    -0.60
    ozy
    -0.59
    wcs
    -0.58
    cest
    -0.58
    ãĥĹ
    -0.58
     tabl
    -0.57
     iceberg
    -0.57
    POSITIVE LOGITS
     ."
    1.32
     .)
    1.02
     .
    1.01
     ..
    0.96
    -.
    0.90
    shall
    0.89
     ,"
    0.87
    _.
    0.86
    `.
    0.84
    shut
    0.83
    Activations Density 0.026%

    No Known Activations