© 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. 24-GEMMASCOPE-2-RES-16K
    7. 8299
    Prev
    Next
    INDEX
    Explanations

    append, add, push

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

    The neuron fires most strongly on imperative or instructional “action” verbs (like “add,” “allocates,” “add to,” etc.) that appear in procedural how-to 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
     perverse
    0.52
     Wasch
    0.52
    准
    0.52
     Vold
    0.51
     Daven
    0.51
     aberrations
    0.50
    是被
    0.50
    čno
    0.49
    prove
    0.49
    程度
    0.49
    POSITIVE LOGITS
     новый
    0.93
    新建
    0.90
    newItem
    0.89
    addItem
    0.86
     новым
    0.84
    ใหม่
    0.83
     الجديد
    0.83
     새
    0.83
     新
    0.82
     ใหม่
    0.82
    Activations Density 0.063%

    No Known Activations