© 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. 23994
    Prev
    Next
    INDEX
    Explanations

    conjunctions and articles, particularly "and" and "the."

    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
    393
    +0.12
    0.6%
    617
    +0.11
    0.5%
    481
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    484
    +0.12
    0.06
    703
    +0.11
    0.05
    393
    +0.11
    0.06
    Negative Logits
    uel
    -0.79
    ali
    -0.71
    ulner
    -0.71
    enezuel
    -0.71
    BUR
    -0.70
    Egypt
    -0.64
    Christ
    -0.63
    Europe
    -0.61
    Brazil
    -0.61
    ANC
    -0.61
    POSITIVE LOGITS
     its
    1.03
     accompanying
    0.89
     thereby
    0.86
     subsequent
    0.85
     ITS
    0.84
     contents
    0.84
     thus
    0.84
     consequently
    0.83
     namesake
    0.83
     skyline
    0.78
    Activations Density 0.262%

    No Known Activations