© 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. 3318
    Prev
    Next
    INDEX
    Explanations

    a jaguar to otro

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

    The neuron fires strongly on attribute-describing noun phrases—especially hyphenated compounds or measurement expressions (e.g. “leadership-related modules,” “Full Size,” “meters” following a numeric value).

    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
    im
    0.89
    it
    0.77
    m
    0.77
    BA
    0.75
    as
    0.72
    ST
    0.72
    UN
    0.71
    F
    0.71
    b
    0.70
    d
    0.69
    POSITIVE LOGITS
    lytres
    0.99
     Brésil
    0.95
     brasile
    0.93
     pantai
    0.93
    цька
    0.92
     pemain
    0.91
     británico
    0.89
     penggunaan
    0.87
     karşılaş
    0.87
     navigateur
    0.86
    Activations Density 0.001%

    No Known Activations