© 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. 12283
    Prev
    Next
    INDEX
    Explanations

    words related to the size, extent, or level of something, especially in a societal or geopolitical context

    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
    188
    +0.11
    0.5%
    616
    +0.10
    0.4%
    160
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    225
    +0.11
    0.06
    188
    +0.10
    0.05
    160
    +0.09
    0.06
    Negative Logits
    ategor
    -0.66
    ocally
    -0.62
     Advertisement
    -0.60
    olkien
    -0.60
    enegger
    -0.60
     Ago
    -0.58
    issy
    -0.58
    inis
    -0.57
    isse
    -0.56
    romeda
    -0.56
    POSITIVE LOGITS
     thereof
    1.14
     of
    1.10
    iest
    0.99
    fulness
    0.88
     inherent
    0.87
    of
    0.84
    lessness
    0.80
    iness
    0.80
     OF
    0.80
    lihood
    0.76
    Activations Density 0.259%

    No Known Activations