© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 15413
    Prev
    Next
    INDEX
    Explanations

    words related to different types of alcoholic beverages, particularly beer

    oai_token-act-pair · gpt-3.5-turbo
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    459
    +0.16
    0.7%
    35
    +0.10
    0.5%
    203
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    459
    +0.16
    0.03
    634
    +0.10
    0.02
    203
    +0.10
    0.02
    Negative Logits
     Telecommunications
    -0.88
     Telecom
    -0.76
    itives
    -0.74
    ortex
    -0.71
    heed
    -0.69
    UTC
    -0.68
    aido
    -0.67
     Lanka
    -0.67
     metic
    -0.66
    inosaur
    -0.63
    POSITIVE LOGITS
     brewed
    1.34
     drinkers
    1.11
     bottles
    0.97
     tasting
    0.94
     cans
    0.92
     bottle
    0.91
     brewing
    0.91
     brewer
    0.90
     brewers
    0.88
     fermented
    0.88
    Activations Density 0.022%

    No Known Activations