© 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. 5322
    Prev
    Next
    INDEX
    Explanations

    issues related to prominence, hierarchy, or importance, particularly focusing on larger-scale events or situations

    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
    487
    +0.12
    0.6%
    534
    +0.10
    0.4%
    539
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    623
    +0.12
    0.04
    669
    +0.10
    0.04
    422
    +0.08
    0.04
    Negative Logits
    dain
    -0.73
    amera
    -0.72
    uilding
    -0.69
    cedes
    -0.64
    Cola
    -0.63
    VERTISEMENT
    -0.63
    iffe
    -0.61
    zyk
    -0.61
    ecause
    -0.61
    SHIP
    -0.60
    POSITIVE LOGITS
    est
    1.19
    liest
    1.09
    iest
    1.04
    hest
    0.99
     ones
    0.95
     confines
    0.85
     stuff
    0.80
     side
    0.80
    most
    0.80
     end
    0.78
    Activations Density 0.222%

    No Known Activations