© 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. 3185
    Prev
    Next
    INDEX
    Explanations

    broken down into

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

    The neuron detects adverbial particles (like up, out, in, over) when they serve as part of a phrasal verb.

    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
     up
    0.93
    Y
    0.76
     lên
    0.74
    之为
    0.72
    t
    0.72
    ขึ้น
    0.70
    出来的
    0.70
    K
    0.68
    ออก
    0.67
    起来
    0.66
    POSITIVE LOGITS
    នូវ
    0.84
     nicely
    0.78
     goodies
    0.73
     materialism
    0.73
     some
    0.70
    一些
    0.70
    了一些
    0.70
    ackerel
    0.70
    нець
    0.70
     questionnaires
    0.69
    Activations Density 0.739%

    No Known Activations