© 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. 7823
    Prev
    Next
    INDEX
    Explanations

    This text introduces

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

    The neuron is strongly activated by numeric tokens—particularly standalone digits, multi‐digit numbers (like years), and related punctuation—i.e. it detects numeral references in the text.

    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
     besides
    0.71
     like
    0.66
     since
    0.61
     nors
    0.59
     probs
    0.59
     huge
    0.59
    及时
    0.57
     maybe
    0.56
     protrusion
    0.56
     etc
    0.55
    POSITIVE LOGITS
    ﺎ
    0.77
     あなた
    0.77
    uitse
    0.76
    ྵ
    0.76
     Очень
    0.75
    Очень
    0.73
    UMB
    0.73
    JL
    0.72
     Multicultural
    0.72
    URED
    0.71
    Activations Density 0.000%

    No Known Activations