© 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. 1223
    Prev
    Next
    INDEX
    Explanations

    childhood

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

    The neuron strongly activates on central, content-bearing nouns (e.g. “base,” “simplicity”)—that is, it picks out high-level topic or theme nouns 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
    「
    0.67
     cũng
    0.65
    本
    0.57
     noticeably
    0.56
    及
    0.55
    滗
    0.55
    （
    0.54
    量
    0.54
    the
    0.54
    #
    0.54
    POSITIVE LOGITS
    PMorgan
    0.62
     فقال
    0.59
    weeney
    0.58
     establecidos
    0.58
     Aussage
    0.58
     amelyet
    0.56
     delanter
    0.55
    รถ
    0.54
     {},
    0.54
    മ്പോ
    0.53
    Activations Density 0.000%

    No Known Activations