© 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 1536
    9. 8-RES_FS1536-JB
    11. 1103
    Prev
    Next
    INDEX
    Explanations

    locations and organizations in Washington, DC

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    254
    +0.13
    0.6%
    23
    +0.11
    0.5%
    339
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    752
    +0.13
    0.07
    200
    +0.11
    0.06
    254
    +0.10
    0.06
    Negative Logits
    leness
    -0.79
    vana
    -0.75
    selection
    -0.71
    ONY
    -0.67
     Cameroon
    -0.66
    lihood
    -0.65
    ensed
    -0.63
    Scroll
    -0.63
    otal
    -0.62
    Integer
    -0.62
    POSITIVE LOGITS
    sonian
    1.17
     DC
    1.16
     Post
    1.02
     Dull
    0.97
     Capitals
    0.97
     Irving
    0.94
     Heights
    0.90
     Redskins
    0.90
     Examiner
    0.87
     Wizards
    0.86
    Activations Density 1.113%

    No Known Activations