© 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. 14998
    Prev
    Next
    INDEX
    Explanations

    phrases indicating a scientific or technical discussion related to specific fields like biology, civilization, and construction

    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
    480
    +0.37
    1.9%
    447
    +0.25
    1.3%
    87
    +0.18
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    480
    +0.37
    0.02
    266
    +0.25
    0.06
    362
    +0.18
    0.07
    Negative Logits
    jri
    -0.63
     destro
    -0.61
     advoc
    -0.60
    Ire
    -0.60
     Vaugh
    -0.60
    aturdays
    -0.60
    ulla
    -0.58
    efully
    -0.58
    wcs
    -0.57
    daq
    -0.56
    POSITIVE LOGITS
     largeDownload
    0.70
     âĵĺ
    0.58
    ¶
    0.56
     Aviation
    0.55
     Abandon
    0.54
     Artificial
    0.54
     Allaah
    0.54
     Xperia
    0.51
     FIG
    0.51
     invention
    0.51
    Activations Density 0.406%

    No Known Activations