© 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. 6679
    Prev
    Next
    INDEX
    Explanations

    technical information and details related to the manufacturing and adoption of specific products, specifically in the context of articles discussing various aspects of the products such as development, popularity, and competition

    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
    447
    +0.22
    1.0%
    481
    +0.14
    0.6%
    309
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.22
    0.06
    249
    +0.14
    0.05
    225
    +0.12
    0.05
    Negative Logits
    $.
    -0.76
    +.
    -0.75
    ().
    -0.72
    _.
    -0.69
    !.
    -0.69
     hereafter
    -0.64
     respectively
    -0.62
    %.
    -0.60
     thirds
    -0.59
    gil
    -0.58
    POSITIVE LOGITS
     lends
    0.92
     afforded
    0.87
     embold
    0.86
     coincided
    0.84
     persists
    0.84
     reflects
    0.83
     ensures
    0.83
     reinforces
    0.83
     seems
    0.82
    lessness
    0.81
    Activations Density 0.320%

    No Known Activations