© 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. 24448
    Prev
    Next
    INDEX
    Explanations

    phrases related to advocacy or pushing for change

    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
    282
    +0.11
    0.5%
    593
    +0.10
    0.5%
    665
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    282
    +0.11
    0.03
    665
    +0.10
    0.03
    50
    +0.10
    0.02
    Negative Logits
    lys
    -0.72
    psons
    -0.69
    atural
    -0.68
     redes
    -0.64
    nam
    -0.63
    omial
    -0.62
    brance
    -0.62
    obyl
    -0.61
    ãĥ¼ãĥĨãĤ£
    -0.61
     Recogn
    -0.60
    POSITIVE LOGITS
     forward
    1.06
    chairs
    0.98
     boundaries
    0.98
     toward
    0.95
     towards
    0.90
     ahead
    0.89
    back
    0.87
     aside
    0.87
     onward
    0.87
     harder
    0.86
    Activations Density 0.035%

    No Known Activations