© 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. 24-GEMMASCOPE-2-RES-16K
    7. 12840
    Prev
    Next
    INDEX
    Explanations

    math notation, programming, or lists

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

    This neuron fires strongly on the very first “real” content word after a segment or paragraph boundary, essentially marking the opening token of each new text chunk.

    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_24_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
    hattisgarh
    0.92
    ز
    0.83
    ל
    0.82
    <0xA2>
    0.82
    ی
    0.82
    ्ड
    0.80
    ת
    0.80
    ະ
    0.80
    دا
    0.79
     hasattr
    0.79
    POSITIVE LOGITS
     tand
    0.70
     avent
    0.66
     lati
    0.66
     grot
    0.65
    sect
    0.64
     dins
    0.63
     pattes
    0.63
     rand
    0.62
     comm
    0.62
    stituting
    0.61
    Activations Density 0.000%

    No Known Activations