© 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. 24478
    Prev
    Next
    INDEX
    Explanations

    references to the Times newspaper

    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
    718
    +0.10
    0.5%
    504
    +0.10
    0.4%
    267
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    475
    +0.10
    0.02
    267
    +0.10
    0.02
    586
    +0.09
    0.02
    Negative Logits
    rd
    -0.94
    rous
    -0.82
    razil
    -0.82
    leased
    -0.78
    isable
    -0.77
    tesy
    -0.75
    icides
    -0.73
    dyl
    -0.73
    atical
    -0.70
    ded
    -0.70
    POSITIVE LOGITS
     Literary
    1.08
     Newsp
    1.03
     Magazine
    0.94
     Dispatch
    0.93
     Square
    0.92
     Herald
    0.86
     bestselling
    0.85
     columnist
    0.85
     Publishing
    0.84
     Reader
    0.83
    Activations Density 0.020%

    No Known Activations