© 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. 24391
    Prev
    Next
    INDEX
    Explanations

    references to the Green Bay Packers

    oai_token-act-pair · gpt-4o-miniTriggered 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
    459
    +0.17
    0.8%
    269
    +0.11
    0.5%
    70
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    459
    +0.17
    0.03
    188
    +0.11
    0.03
    269
    +0.11
    0.02
    Negative Logits
    nces
    -0.79
    orically
    -0.78
    Downloadha
    -0.73
    gerald
    -0.68
    Reloaded
    -0.67
    ysis
    -0.66
    ymes
    -0.66
    umbn
    -0.64
    ebin
    -0.64
    oldemort
    -0.62
    POSITIVE LOGITS
    grass
    1.18
    wich
    1.04
    baum
    1.04
    leaf
    1.04
     Lantern
    0.94
    field
    0.93
    house
    0.93
    lit
    0.92
    stuff
    0.91
    lee
    0.90
    Activations Density 0.021%

    No Known Activations