© 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. 20906
    Prev
    Next
    INDEX
    Explanations

    personal beliefs and opinions

    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
    566
    +0.14
    0.6%
    718
    +0.13
    0.6%
    697
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    502
    +0.14
    0.09
    566
    +0.13
    0.09
    354
    +0.12
    0.09
    Negative Logits
     Reverse
    -0.65
    tnc
    -0.63
    pires
    -0.62
    ses
    -0.60
     Gap
    -0.57
     undisclosed
    -0.57
    ãĥ¬
    -0.56
    interstitial
    -0.56
     prosecution
    -0.56
    groupon
    -0.56
    POSITIVE LOGITS
    'm
    1.37
    've
    1.28
     dunno
    1.14
    'll
    1.14
     suppose
    1.12
    'd
    1.10
    ronic
    1.03
     mean
    1.00
     guess
    0.98
     wonder
    0.98
    Activations Density 0.197%

    No Known Activations