© 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. 73154
    Prev
    Next
    INDEX
    Explanations

    code, questions, foreign words

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

    The neuron strongly activates on programming loop constructs (e.g. “for”, “do”, “while” and their enclosing parentheses/braces), effectively detecting occurrences of loops in code.

    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
    是我的
    -0.94
    az
    -0.90
    ap
    -0.85
     from
    -0.83
    aggio
    -0.82
    onaceous
    -0.82
    acar
    -0.82
    なと
    -0.79
    Any
    -0.78
     their
    -0.77
    POSITIVE LOGITS
     esetén
    0.85
    Trotz
    0.84
    Aunque
    0.84
    Angeb
    0.84
     NOTICE
    0.82
    koc
    0.82
    necessarily
    0.82
    女神
    0.82
     названия
    0.82
    Schul
    0.81
    Activations Density 0.001%

    No Known Activations