© 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. 22761
    Prev
    Next
    INDEX
    Explanations

    structured elements like sections, headings, or labels that indicate organization or categorization within the text

    oai_token-act-pair · gpt-4o-miniTriggered 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
    481
    +0.16
    0.7%
    333
    +0.11
    0.5%
    609
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    609
    +0.16
    0.04
    288
    +0.11
    0.04
    136
    +0.10
    0.03
    Negative Logits
    Downloadha
    -0.78
     challeng
    -0.73
    undai
    -0.72
    .","
    -0.70
    creen
    -0.67
    avorite
    -0.66
    cale
    -0.66
    merce
    -0.63
    perty
    -0.63
    heet
    -0.62
    POSITIVE LOGITS
    wise
    1.02
     wise
    0.96
    ually
    0.93
     Explan
    0.80
    ual
    0.76
     Facts
    0.75
     Details
    0.73
    atically
    0.71
    ally
    0.71
     Size
    0.71
    Activations Density 0.205%

    No Known Activations