© 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 24576
    9. 8-RES_FS24576-JB
    11. 7932
    Prev
    Next
    INDEX
    Explanations

    general statements or overviews in text

    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
    450
    +0.14
    0.6%
    709
    +0.10
    0.4%
    483
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    450
    +0.14
    0.04
    423
    +0.10
    0.04
    231
    +0.09
    0.04
    Negative Logits
    aily
    -0.73
    him
    -0.70
    ËĪ
    -0.70
    fest
    -0.66
    imm
    -0.64
    ocaust
    -0.62
    aciously
    -0.61
    ocracy
    -0.60
    gio
    -0.59
    vg
    -0.58
    POSITIVE LOGITS
    adays
    0.98
     speaking
    0.89
    ccording
    0.80
    entimes
    0.77
     Speaking
    0.73
     we
    0.73
     there
    0.73
    terday
    0.73
     though
    0.69
     commenters
    0.69
    Activations Density 0.131%

    No Known Activations