© 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. 22-GEMMASCOPE-RES-131K
    7. 39036
    Prev
    Next
    INDEX
    Explanations

    XML and code structures

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

    The neuron strongly activates on code‐identifier tokens (e.g. class, namespace, or other programming symbols), effectively marking out programming identifiers in the text.

    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
     will
    -0.91
     otro
    -0.84
     салон
    -0.82
    adering
    -0.77
     cabina
    -0.77
    cestors
    -0.77
    遽
    -0.76
     accumulating
    -0.74
     revisa
    -0.74
     Bracelets
    -0.72
    POSITIVE LOGITS
    Triangle
    1.00
    paces
    0.91
    _/
    0.89
    等の
    0.89
    pac
    0.88
    Postcode
    0.88
    tingly
    0.87
    quarter
    0.86
     át
    0.85
    IJ
    0.85
    Activations Density 0.003%

    No Known Activations