© 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 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. 12845
    Prev
    Next
    INDEX
    Explanations

    exclamations with varying intensities

    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
    481
    +0.11
    0.5%
    475
    +0.11
    0.5%
    6
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.11
    0.06
    267
    +0.11
    0.04
    440
    +0.11
    0.03
    Negative Logits
     conduc
    -0.79
    undai
    -0.73
    rament
    -0.71
     destro
    -0.67
     reven
    -0.65
     utilitarian
    -0.64
    ogram
    -0.63
    urai
    -0.62
    senal
    -0.61
    yrus
    -0.60
    POSITIVE LOGITS
    @#&
    1.39
    ?!
    1.22
    #$
    1.18
    ãĢį
    1.05
    @#
    1.03
     :)
    0.98
     :-)
    0.94
     ðŁĺ
    0.94
     Needless
    0.92
     Anyway
    0.92
    Activations Density 0.066%

    No Known Activations