© 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. 4137
    Prev
    Next
    INDEX
    Explanations

    phrase related to 'item' or 'task'

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

    The neuron highly activates on boilerplate–style, promotional call-to-action text (e.g. “Make sure…,” “Do not forget…,” “you…”)—in other words, marketing/website-promo language.

    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
     biopsies
    0.75
     honti
    0.71
    Putting
    0.70
    тивы
    0.69
    ័យ
    0.67
     monolayers
    0.67
     prisión
    0.66
    }|=
    0.66
     karet
    0.65
     radiographs
    0.64
    POSITIVE LOGITS
    apons
    0.83
    ARD
    0.79
     contribuer
    0.77
     Leistung
    0.74
     وك
    0.73
     é
    0.73
    ｴ
    0.73
     Farley
    0.72
    iance
    0.72
     фі
    0.71
    Activations Density 0.119%

    No Known Activations