© 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. 21294
    Prev
    Next
    INDEX
    Explanations

    the word "most" followed by a number

    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
    190
    +0.11
    0.5%
    671
    +0.10
    0.5%
    408
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    190
    +0.11
    0.05
    194
    +0.10
    0.04
    726
    +0.10
    0.04
    Negative Logits
    kas
    -0.81
    æ©
    -0.77
    heid
    -0.77
     Certification
    -0.70
    instead
    -0.70
    fell
    -0.69
    selves
    -0.69
     Chips
    -0.69
    pload
    -0.67
     Films
    -0.66
    POSITIVE LOGITS
     important
    1.17
     obvious
    1.17
     notable
    1.10
     interesting
    1.07
     noticeable
    1.07
     plausible
    1.05
     likely
    1.03
     salient
    1.01
     prominent
    1.00
     glaring
    1.00
    Activations Density 0.042%

    No Known Activations