© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 7959
    Prev
    Next
    INDEX
    Explanations

    This neuron is especially sensitive to action‐oriented verbs (i.e. instruction or recommendation words) appearing 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_31_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
    isations
    0.72
    ার্চ
    0.71
     ошибки
    0.70
    磐
    0.70
     produto
    0.68
    izers
    0.67
    ités
    0.66
     vantagem
    0.66
    чее
    0.66
    guessed
    0.65
    POSITIVE LOGITS
    นาย
    0.82
    aa
    0.75
    s
    0.72
     imong
    0.71
     which
    0.70
     accompanied
    0.70
    yll
    0.70
    ที่คุณ
    0.69
     as
    0.68
    ska
    0.68
    Activations Density 0.000%

    No Known Activations