© 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. 17242
    Prev
    Next
    INDEX
    Explanations

    information related to ongoing investigations, legal proceedings, and criticisms

    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
    55
    +0.17
    0.8%
    322
    +0.09
    0.4%
    79
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    55
    +0.17
    0.05
    79
    +0.09
    0.03
    208
    +0.09
    0.03
    Negative Logits
    stice
    -0.79
    ouses
    -0.71
    anish
    -0.70
    article
    -0.65
    cies
    -0.65
    regate
    -0.64
    urches
    -0.64
    oops
    -0.63
    orum
    -0.63
    ision
    -0.63
    POSITIVE LOGITS
     noteworthy
    0.92
     notable
    0.92
     known
    0.89
     considered
    0.88
     able
    0.88
     referred
    0.88
     tasked
    0.87
     implicated
    0.87
     responsible
    0.87
     included
    0.85
    Activations Density 0.067%

    No Known Activations