© 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. 26479
    Prev
    Next
    INDEX
    Explanations

    code assertions and descriptions

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

    This neuron fires most strongly on leading whitespace indentation in code (i.e. runs of spaces at the start of a line).

    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
     PEDRO
    -0.93
    чно
    -0.91
     ceux
    -0.90
    ύ
    -0.86
     می‌تواند
    -0.86
     информация
    -0.85
    если
    -0.84
    ewoo
    -0.83
     بعضی
    -0.83
     Если
    -0.82
    POSITIVE LOGITS
     any
    1.08
     Sortiment
    1.02
     ^{\
    0.96
     in
    0.92
    LES
    0.91
    ORTH
    0.91
    stateMutability
    0.91
     桃
    0.91
    СТВА
    0.90
     volverá
    0.88
    Activations Density 0.007%

    No Known Activations