© 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. 4179
    Prev
    Next
    INDEX
    Explanations

    quotations or emphasized phrases denoting strong feelings or opinions

    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
    336
    +0.10
    0.5%
    537
    +0.10
    0.4%
    706
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    537
    +0.10
    0.02
    132
    +0.10
    0.02
    517
    +0.09
    0.02
    Negative Logits
    SourceFile
    -0.84
    endor
    -0.80
    nown
    -0.80
    mens
    -0.79
    inav
    -0.78
    isl
    -0.77
    ãĥ¼ãĥ«
    -0.77
    ãĥ´
    -0.76
    manship
    -0.72
    ItemImage
    -0.72
    POSITIVE LOGITS
     motto
    0.63
     verdict
    0.62
     HDR
    0.62
     Fallon
    0.60
     Blackburn
    0.58
     campaigners
    0.58
     slogan
    0.57
     Wynne
    0.57
    ',
    0.56
     hacks
    0.56
    Activations Density 0.035%

    No Known Activations