© 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. 16245
    Prev
    Next
    INDEX
    Explanations

    code syntax and keywords

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

    The neuron fires on code‐style tokens—things like punctuation (e.g. “.”, “–”), uppercase identifiers (e.g. “LIBS”, “Case”), and hyphenated names (e.g. “fa–svg”)—i.e. markers of programming or markup syntax.

    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
     certamente
    -0.95
    もしろ
    -0.94
    kennen
    -0.92
     suporta
    -0.92
     necessariamente
    -0.89
     กระ
    -0.88
     proveedores
    -0.88
     aktiviert
    -0.87
    оне
    -0.86
    ಪ
    -0.86
    POSITIVE LOGITS
     still
    1.14
     even
    0.96
     so
    0.95
    ุทธ
    0.86
     pengaruh
    0.85
    standes
    0.84
     only
    0.81
     todavía
    0.80
     ;
    0.80
     it
    0.80
    Activations Density 0.009%

    No Known Activations