© 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. 3265
    Prev
    Next
    INDEX
    Explanations

    em dash—used for emphasis or to separate information in a sentence

    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
    160
    +0.16
    0.8%
    308
    +0.15
    0.7%
    314
    +0.15
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    308
    +0.16
    0.04
    160
    +0.15
    0.05
    546
    +0.15
    0.03
    Negative Logits
     palm
    -0.77
     oven
    -0.75
     mail
    -0.72
     barr
    -0.71
     mum
    -0.71
     indemn
    -0.70
     nuts
    -0.69
     handic
    -0.69
     pronoun
    -0.68
     nut
    -0.67
    POSITIVE LOGITS
    meaning
    1.37
    perhaps
    1.33
    that
    1.32
    especially
    1.31
    which
    1.31
    these
    1.29
    including
    1.29
    particularly
    1.28
    they
    1.28
    albeit
    1.27
    Activations Density 0.054%

    No Known Activations