© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 15109
    Prev
    Next
    INDEX
    Explanations

    respectful, healthy, safe discussions

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

    This neuron flags individual high-salience content words—uncommon or informative terms (e.g. “pilot,” “nutshell,” “bid,” “certain”) that carry strong lexical weight 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
    ients
    0.68
    iton
    0.68
    并且
    0.66
    arkar
    0.65
    ই
    0.64
    न
    0.64
    voud
    0.64
    itud
    0.63
    và
    0.62
     dagen
    0.61
    POSITIVE LOGITS
     nutshell
    1.03
     найбільш
    0.74
    㸵
    0.68
    sequ
    0.66
     hurry
    0.65
     constituency
    0.65
    clockwise
    0.65
    cont
    0.65
    Seperti
    0.64
     exclusionary
    0.63
    Activations Density 0.071%

    No Known Activations