© 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 EvalsExportsinterp-engineNEWAPI 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. 16438
    Prev
    Next
    INDEX
    Explanations

    information related to statistical data and research findings, particularly focusing on figures and averages

    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
    592
    +0.11
    0.5%
    208
    +0.10
    0.5%
    102
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    102
    +0.11
    0.06
    208
    +0.10
    0.04
    55
    +0.10
    0.05
    Negative Logits
    ossom
    -0.78
     refrain
    -0.72
     Awakens
    -0.71
    iate
    -0.68
    wake
    -0.67
    heny
    -0.66
    inav
    -0.66
    ventures
    -0.66
     launches
    -0.64
     wills
    -0.64
    POSITIVE LOGITS
     estimated
    1.06
     negligible
    0.97
     considerably
    0.92
     slashed
    0.87
     reduced
    0.87
     doubled
    0.83
     computed
    0.83
     pegged
    0.83
     unchanged
    0.81
     ¥
    0.81
    Activations Density 0.174%

    No Known Activations