© 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. 5259
    Prev
    Next
    INDEX
    Explanations

    information related to news articles, including dates, events, locations, and legal proceedings

    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
    87
    +0.31
    1.6%
    373
    +0.14
    0.8%
    752
    +0.09
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    87
    +0.31
    -0.03
    711
    +0.14
    0.03
    516
    +0.09
    0.04
    Negative Logits
    âĢ¦)
    -0.57
    ...)
    -0.55
     Appears
    -0.52
     HUGE
    -0.51
     LOT
    -0.51
    unless
    -0.50
     partName
    -0.49
     fascinating
    -0.48
    whatever
    -0.47
     âĵĺ
    -0.47
    POSITIVE LOGITS
     unsuccessfully
    0.59
     improperly
    0.59
    itored
    0.58
     earlier
    0.58
     cumbers
    0.56
     unexpectedly
    0.55
     allegedly
    0.53
     unsuccessful
    0.53
     prematurely
    0.53
     previous
    0.52
    Activations Density 1.090%

    No Known Activations