© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 7981
    Prev
    Next
    INDEX
    Explanations

    the word 'cept'

    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
    339
    +0.12
    0.5%
    314
    +0.11
    0.5%
    289
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.12
    0.03
    289
    +0.11
    0.01
    326
    +0.10
    0.02
    Negative Logits
     grandparents
    -0.68
     haste
    -0.67
     intest
    -0.66
     iTunes
    -0.66
    ups
    -0.64
     dining
    -0.63
     grandfather
    -0.63
     slaughtered
    -0.63
     descent
    -0.62
     segments
    -0.62
    POSITIVE LOGITS
    ible
    1.39
    ics
    1.34
    icism
    1.30
    icons
    1.30
    ibly
    1.26
    ical
    1.26
    icon
    1.24
    oric
    1.21
    ually
    1.19
    ual
    1.18
    Activations Density 0.056%

    No Known Activations