© 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. 7518
    Prev
    Next
    INDEX
    Explanations

    phrases emphasizing positive qualities or comparison to others

    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
    373
    +0.13
    0.6%
    689
    +0.11
    0.5%
    557
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    13
    +0.13
    0.03
    389
    +0.11
    0.02
    408
    +0.09
    0.03
    Negative Logits
    giene
    -0.75
    edom
    -0.72
    NetMessage
    -0.69
    igham
    -0.69
    hoe
    -0.66
    naires
    -0.66
    edia
    -0.65
    ities
    -0.65
    isk
    -0.65
    wick
    -0.63
    POSITIVE LOGITS
     testament
    0.78
     illustrating
    0.72
     than
    0.72
    =>
    0.68
    âĢ¢âĢ¢âĢ¢âĢ¢
    0.63
     compliments
    0.63
     linem
    0.63
     exempl
    0.62
    TPPStreamerBot
    0.62
     thrilled
    0.61
    Activations Density 0.087%

    No Known Activations