© 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. 24-GEMMASCOPE-2-RES-16K
    7. 3020
    Prev
    Next
    INDEX
    Explanations

    list items, numbers, and names

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

    The neuron is specifically sensitive to numeric tokens (particularly multi‐digit numbers and decimal numbers).

    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
     utilizz
    0.63
     inextric
    0.63
    
    0.62
     attacker
    0.61
     substituents
    0.61
     subjekt
    0.61
     irrevoc
    0.60
    ѣ
    0.60
     вже
    0.59
     inextricably
    0.59
    POSITIVE LOGITS
    T
    0.81
    The
    0.76
    a
    0.75
    An
    0.72
    A
    0.72
    L
    0.72
    J
    0.72
    C
    0.71
    an
    0.71
    Thomas
    0.71
    Activations Density 0.014%

    No Known Activations