© 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. 4631
    Prev
    Next
    INDEX
    Explanations

    words related to sizes or quantities, appearing in various contexts

    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
    314
    +0.11
    0.5%
    537
    +0.11
    0.5%
    383
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    537
    +0.11
    0.01
    89
    +0.11
    0.01
    232
    +0.09
    0.01
    Negative Logits
     Danger
    -0.71
     Compass
    -0.68
     Dreams
    -0.65
     reforming
    -0.65
    eering
    -0.64
    Ranked
    -0.62
     Pru
    -0.59
     compe
    -0.57
     Freedom
    -0.57
     improving
    -0.57
    POSITIVE LOGITS
    iatures
    1.43
    otaur
    1.35
    nesota
    1.31
    utes
    1.18
    neapolis
    1.18
    istered
    1.13
    usc
    1.06
    isite
    1.05
    erva
    1.03
    ivan
    1.01
    Activations Density 0.009%

    No Known Activations