© 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. 22827
    Prev
    Next
    INDEX
    Explanations

    the presence and frequency of the '&' character in the text, which likely indicates specific formatting or parameters in a coding or data context

    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
    160
    +0.12
    0.6%
    290
    +0.12
    0.5%
    465
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.12
    0.02
    0
    +0.12
    0.01
    107
    +0.11
    0.01
    Negative Logits
    issance
    -0.92
     Citiz
    -0.76
    untled
    -0.68
     dispers
    -0.67
     Borg
    -0.66
     Oblivion
    -0.65
    ilty
    -0.65
     plumbing
    -0.65
    ivating
    -0.65
    ifying
    -0.64
    POSITIVE LOGITS
    utm
    1.09
    hl
    1.05
    gt
    1.00
    lt
    0.88
    cas
    0.87
    SOURCE
    0.87
    optional
    0.86
    amp
    0.82
    ãĤ´ãĥ³
    0.80
    hei
    0.80
    Activations Density 0.007%

    No Known Activations