© 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 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. 6450
    Prev
    Next
    INDEX
    Explanations

    mentions of political events or developments

    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
    +0.39
    1.9%
    470
    +0.12
    0.6%
    373
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    470
    +0.39
    0.04
    157
    +0.12
    0.06
    480
    +0.08
    -0.01
    Negative Logits
     thereof
    -0.72
     etc
    -0.68
     [&
    -0.66
    .[
    -0.65
    âĢ¦]
    -0.65
    ."[
    -0.64
    Zip
    -0.63
    $.
    -0.62
    Liter
    -0.61
    CONCLUS
    -0.61
    POSITIVE LOGITS
     controvers
    1.00
     plagued
    0.83
     notoriously
    0.83
     accused
    0.81
     fiercely
    0.81
     dogged
    0.80
     embroiled
    0.80
     criticised
    0.80
     ousted
    0.79
     controversial
    0.78
    Activations Density 1.106%

    No Known Activations