© 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. 12494
    Prev
    Next
    INDEX
    Explanations

    code snippets related to coding languages

    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
    387
    +0.12
    0.5%
    745
    +0.11
    0.5%
    230
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.12
    0.03
    284
    +0.11
    0.04
    153
    +0.11
    0.03
    Negative Logits
    letter
    -0.84
     Dion
    -0.81
    Wing
    -0.77
     toes
    -0.74
    hor
    -0.72
     Flan
    -0.71
     Corona
    -0.71
     HOR
    -0.70
     Laur
    -0.70
     Thy
    -0.69
    POSITIVE LOGITS
    00
    0.94
    itt
    0.94
     00
    0.92
    003
    0.89
    299
    0.87
    usted
    0.87
    itely
    0.87
     299
    0.79
    009
    0.77
    ff
    0.77
    Activations Density 0.213%

    No Known Activations