© 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. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 15177
    Prev
    Next
    INDEX
    Explanations

    code response package

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

    The neuron fires strongly on programming language type‐declaration tokens—e.g. class, interface/struct names and related keywords—marking where new types are defined.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     even
    -1.13
     for
    -1.08
    &
    -1.06
    tative
    -1.05
     at
    -1.04
     in
    -1.02
     tropicales
    -1.02
    =?,
    -1.02
     newOwner
    -0.96
    gallon
    -0.96
    POSITIVE LOGITS
    mitives
    1.00
     appeler
    0.99
    結局
    0.99
     énorm
    0.99
     laiss
    0.98
    咻
    0.97
    ormais
    0.97
    綜合報導
    0.96
     traités
    0.95
    の詳細
    0.94
    Activations Density 0.006%

    No Known Activations