© 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. 2789
    Prev
    Next
    INDEX
    Explanations

    parentheses, brackets, and colons

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

    The neuron fires on those stand‐out product or policy keywords—tokens that describe a product’s condition (“new”), care instructions (“wash” in “machine wash”), or licensing qualifiers (“non-exclusive,” “non-transferable”).

    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
    য়ান
    0.51
     tambah
    0.48
     misfort
    0.46
    dotnet
    0.46
    拜
    0.46
    toolbar
    0.46
    菏
    0.46
    LeftSide
    0.44
    Println
    0.44
    OfDeath
    0.44
    POSITIVE LOGITS
    ।
    0.67
    。
    0.61
     and
    0.55
    และ
    0.53
     και
    0.52
     but
    0.51
     và
    0.51
     with
    0.50
    .
    0.48
     ও
    0.47
    Activations Density 0.002%

    No Known Activations