© 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. 12333
    Prev
    Next
    INDEX
    Explanations

    features and benefits of a product or service described in a technical manner

    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
    640
    +0.25
    1.2%
    481
    +0.21
    1.1%
    160
    +0.18
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.25
    0.08
    160
    +0.21
    0.09
    107
    +0.18
    0.06
    Negative Logits
     today
    -0.76
     mund
    -0.73
     curly
    -0.72
     tree
    -0.72
     gir
    -0.70
     trump
    -0.70
     wagon
    -0.69
     Cabrera
    -0.69
     Shelby
    -0.69
     camp
    -0.68
    POSITIVE LOGITS
    Avoid
    1.48
    Conclusion
    1.43
    License
    1.37
    Particip
    1.35
    Increased
    1.34
    Detailed
    1.34
    Various
    1.34
    Includes
    1.33
    Contribut
    1.33
    Use
    1.33
    Activations Density 0.219%

    No Known Activations