© 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. 1891
    Prev
    Next
    INDEX
    Explanations

    names of different people and plants

    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
    87
    +0.41
    2.5%
    314
    +0.11
    0.7%
    176
    +0.08
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.41
    0.09
    232
    +0.11
    0.08
    314
    +0.08
    0.01
    Negative Logits
     cryptocurrency
    -0.53
     Patreon
    -0.48
     Reloaded
    -0.46
     Canaver
    -0.46
     Gamergate
    -0.46
     Libertarian
    -0.45
     blockchain
    -0.45
     worthwhile
    -0.45
     mandate
    -0.45
     Perspective
    -0.45
    POSITIVE LOGITS
    ona
    0.64
    itto
    0.63
    zes
    0.60
    tek
    0.60
    ola
    0.60
    acia
    0.60
    areth
    0.58
    ango
    0.57
    ito
    0.57
    obar
    0.56
    Activations Density 0.585%

    No Known Activations