© 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 EvalsExportsinterp-engineNEWAPI 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. 2845
    Prev
    Next
    INDEX
    Explanations

    words related to the DC Comics franchise

    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
    314
    +0.11
    0.5%
    323
    +0.11
    0.5%
    527
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    527
    +0.11
    0.01
    508
    +0.11
    0.01
    323
    +0.11
    0.01
    Negative Logits
    lihood
    -1.01
    xual
    -0.89
    htar
    -0.77
    tery
    -0.75
    olulu
    -0.75
     htt
    -0.75
    zai
    -0.72
     Finnish
    -0.71
    llor
    -0.71
    pects
    -0.67
    POSITIVE LOGITS
     Comics
    1.22
    NF
    1.16
    MJ
    0.87
    AT
    0.86
    ADA
    0.85
    NM
    0.82
    Leaks
    0.82
    DC
    0.79
    MC
    0.79
     Universe
    0.78
    Activations Density 0.017%

    No Known Activations