© 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. 7917
    Prev
    Next
    INDEX
    Explanations

    statements related to political and economic issues

    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
    483
    +0.10
    0.5%
    56
    +0.10
    0.4%
    752
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    120
    +0.10
    0.04
    537
    +0.10
    0.04
    396
    +0.10
    0.03
    Negative Logits
    swick
    -0.80
    ombies
    -0.77
     oranges
    -0.75
    visors
    -0.73
    irez
    -0.71
     Balls
    -0.70
    lights
    -0.70
     Twins
    -0.70
    loads
    -0.70
     Surge
    -0.69
    POSITIVE LOGITS
     economical
    1.10
     linguistic
    0.99
    technical
    0.99
     cultural
    0.97
     logistical
    0.95
     psychological
    0.95
     philosophical
    0.95
     interpersonal
    0.95
     ecological
    0.95
     behavioural
    0.94
    Activations Density 0.122%

    No Known Activations