© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 108896
    Prev
    Next
    INDEX
    Explanations

    code constructs

    np_acts-logits-general · gemini-2.5-flash-lite

    This neuron fires on code punctuation—especially colons and parentheses that show up in programming‐language constructs.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    onesia
    -0.89
    archie
    -0.88
    thun
    -0.84
    intendo
    -0.82
     Think
    -0.82
    chait
    -0.82
    vaz
    -0.81
    特に
    -0.79
    wiat
    -0.77
    frig
    -0.77
    POSITIVE LOGITS
     _
    1.62
     its
    1.44
     &_
    1.23
     this
    1.12
     our
    1.11
     \_
    1.10
     *_
    1.10
    :_
    1.09
    ->_
    1.02
    -_
    0.99
    Activations Density 0.059%

    No Known Activations