© 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. 3031
    Prev
    Next
    INDEX
    Explanations

    compressed wood chips

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

    The neuron fires on adverbs and modal/qualifier words (e.g. “possible,” “maybe,” “really,” “exactly”) that mark hedges or emphatic qualifiers.

    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
     dello
    0.73
     dye
    0.71
     നിങ്ങളുടെ
    0.70
     sew
    0.69
     ชื่อ
    0.69
    다는
    0.67
    wej
    0.66
    ,}$
    0.65
    .")
    0.64
    다고
    0.64
    POSITIVE LOGITS
    Griff
    0.78
     那麼
    0.72
    ాలను
    0.71
     Griff
    0.70
    दक्षि
    0.69
    arrage
    0.66
    Coconut
    0.65
    aguars
    0.65
     कमलेश
    0.64
     Maugin
    0.64
    Activations Density 0.069%

    No Known Activations