© 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 EvalsExportsAPI 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. 18159
    Prev
    Next
    INDEX
    Explanations

    the word "lightning"

    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
    421
    +0.10
    0.5%
    85
    +0.09
    0.4%
    549
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.10
    0.01
    538
    +0.09
    0.01
    421
    +0.09
    0.01
    Negative Logits
    glers
    -0.94
    ablishment
    -0.84
    yrinth
    -0.83
    eday
    -0.81
    nee
    -0.78
    udo
    -0.78
    rador
    -0.77
    ester
    -0.77
    alse
    -0.76
    orget
    -0.75
    POSITIVE LOGITS
    bolt
    1.19
     bolt
    1.11
     bolts
    1.02
     Bolt
    0.95
     ray
    0.88
     rod
    0.86
    storms
    0.85
     rays
    0.84
     strike
    0.82
    storm
    0.81
    Activations Density 0.023%

    No Known Activations