© 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. 20864
    Prev
    Next
    INDEX
    Explanations

    proper nouns and names, particularly related to individuals

    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
    314
    +0.21
    1.0%
    640
    +0.15
    0.7%
    373
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    627
    +0.21
    0.05
    160
    +0.15
    0.06
    314
    +0.11
    0.02
    Negative Logits
     entitle
    -0.74
     habitable
    -0.74
     eaves
    -0.73
     retard
    -0.70
     guarantee
    -0.70
     exporting
    -0.69
     commonplace
    -0.69
     possible
    -0.68
     lull
    -0.68
     nightly
    -0.67
    POSITIVE LOGITS
    pictured
    1.57
    pron
    1.42
    aka
    1.41
    formerly
    1.39
    whose
    1.24
    via
    1.20
    also
    1.15
    sic
    1.15
    nee
    1.14
    who
    1.09
    Activations Density 0.093%

    No Known Activations