© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 9587
    Prev
    Next
    INDEX
    Explanations

    variables with names that may be misleading or hard to understand

    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
    623
    +0.12
    0.6%
    739
    +0.11
    0.5%
    274
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    623
    +0.12
    0.01
    540
    +0.11
    0.01
    255
    +0.10
    0.01
    Negative Logits
    ļé
    -0.88
    llah
    -0.87
    Ķ
    -0.86
    İ
    -0.83
    lyak
    -0.83
    Ĥª
    -0.81
    lez
    -0.80
    furt
    -0.77
    Fal
    -0.74
    udeb
    -0.74
    POSITIVE LOGITS
     variables
    0.89
    ially
    0.84
    variable
    0.79
     variable
    0.78
     marker
    0.69
     markers
    0.69
    wine
    0.68
     Registrar
    0.68
     names
    0.68
    izations
    0.68
    Activations Density 0.025%

    No Known Activations