© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 12999
    Prev
    Next
    INDEX
    Explanations

    think of

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

    The neuron primarily detects numeric tokens—especially decimal or floating‐point numbers used as measurements, statistics, or timestamps.

    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
    聽
    0.78
    間に
    0.75
    sl
    0.74
    1
    0.74
    I
    0.71
    스로
    0.70
    2
    0.69
    OR
    0.67
    Te
    0.67
    Feb
    0.67
    POSITIVE LOGITS
     of
    2.97
     của
    2.45
    ของการ
    2.33
     của
    2.33
     ofthe
    2.20
    ของ
    2.08
     של
    1.98
     των
    1.94
     της
    1.62
    នៃ
    1.56
    Activations Density 2.144%

    No Known Activations