© 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. 10-GEMMASCOPE-RES-131K
    7. 12816
    Prev
    Next
    INDEX
    Explanations

    loop variables like i, j

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

    The neuron fires on code tokens that are simple variable identifiers—especially single‐letter loop counters or similar short names like “r,” “j,” “i,” “u,” etc.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     that
    -2.67
    .”
    -2.08
    which
    -2.05
    .“
    -1.88
    我們
    -1.85
     穴
    -1.77
     of
    -1.76
     mundiales
    -1.76
    gelijke
    -1.74
     pubblicato
    -1.74
    POSITIVE LOGITS
     was
    3.09
     didn
    2.25
     has
    1.95
     wasn
    1.90
     couldn
    1.89
     comprando
    1.80
     usado
    1.79
     одним
    1.78
     poderosa
    1.74
    アクセサリー
    1.74
    Activations Density 0.037%

    No Known Activations