© 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. 2634
    Prev
    Next
    INDEX
    Explanations

    references to fictional magical schools and characters

    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
    167
    +0.13
    0.6%
    288
    +0.13
    0.6%
    521
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.13
    0.02
    48
    +0.13
    0.01
    99
    +0.12
    0.01
    Negative Logits
    uters
    -0.70
    PsyNetMessage
    -0.67
     Yards
    -0.66
     Fault
    -0.64
    DonaldTrump
    -0.63
     Lima
    -0.61
    rontal
    -0.60
    OPLE
    -0.60
     Geo
    -0.59
     Survivors
    -0.59
    POSITIVE LOGITS
    ry
    1.09
    loo
    0.95
    fly
    0.87
    ery
    0.86
    sonian
    0.80
     Gand
    0.77
    boys
    0.76
     apprentice
    0.76
    boy
    0.75
    arium
    0.75
    Activations Density 0.033%

    No Known Activations