© 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. 5329
    Prev
    Next
    INDEX
    Explanations

    For neuron 4, the main thing it does is find cooking instructions and recipes

    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
    57
    +0.13
    0.6%
    36
    +0.12
    0.6%
    659
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.13
    0.02
    57
    +0.12
    0.01
    311
    +0.12
    0.01
    Negative Logits
    lessly
    -0.80
    papers
    -0.76
    tics
    -0.71
    ships
    -0.71
    mercial
    -0.70
    breaks
    -0.67
    姫
    -0.66
     Establishment
    -0.65
    naire
    -0.65
    SHIP
    -0.65
    POSITIVE LOGITS
    igon
    1.07
    pling
    1.04
    vers
    1.00
    plin
    1.00
    plings
    1.00
    Ga
    0.98
    adish
    0.96
    uten
    0.95
    izo
    0.93
    iva
    0.93
    Activations Density 0.009%

    No Known Activations