© 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. 2809
    Prev
    Next
    INDEX
    Explanations

    phrases related to social issues and societal impacts, particularly related to technology, employment, community, and development

    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
    480
    +0.50
    2.8%
    87
    +0.22
    1.2%
    266
    +0.17
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    266
    +0.50
    0.08
    480
    +0.22
    0.03
    288
    +0.17
    0.09
    Negative Logits
    ernels
    -0.78
    sites
    -0.78
    rooms
    -0.77
    eworks
    -0.76
     Vaugh
    -0.73
    assies
    -0.72
    hops
    -0.72
    units
    -0.71
    casts
    -0.71
     Rooms
    -0.71
    POSITIVE LOGITS
     standpoint
    0.69
     statement
    0.68
     version
    0.67
     piece
    0.66
     apology
    0.65
     article
    0.64
     lawsuit
    0.63
     perspective
    0.63
     spokesperson
    0.63
     sample
    0.63
    Activations Density 0.567%

    No Known Activations