© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 10596
    Prev
    Next
    INDEX
    Explanations

    code imports and declarations

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

    The neuron selectively responds to alphanumeric codes—long sequences mixing uppercase letters and digits (e.g., transaction IDs, barcodes, and similar coded fields).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ({...
    0.59
     preventive
    0.56
     prevState
    0.55
    prevState
    0.54
     компо
    0.53
     (
    0.52
     hoisting
    0.51
    Auth
    0.51
     Pons
    0.50
     ({
    0.50
    POSITIVE LOGITS
    рт
    0.61
    率は
    0.58
    Txt
    0.57
    ডিস
    0.57
     വിള
    0.56
     миллиар
    0.56
     dataGridView
    0.55
     }-\
    0.55
     dokładnie
    0.55
     Monro
    0.55
    Activations Density 0.002%

    No Known Activations