© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 5098
    Prev
    Next
    INDEX
    Explanations

    front/before followed by of/the

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

    The neuron strongly activates on two‐word phrasal verbs combining an action verb with a directional particle (e.g. “set … down,” “sat down,” “drag out”).

    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
     
    0.86
     (
    0.82
     or
    0.82
    /
    0.76
     this
    0.75
     Special
    0.73
     [
    0.72
     "
    0.71
     also
    0.71
     and
    0.71
    POSITIVE LOGITS
     комна
    1.03
     habitación
    0.89
     дверь
    0.88
     набор
    0.87
     двер
    0.87
     ноутбу
    0.86
     отправить
    0.85
     приготов
    0.84
    🚪
    0.82
     поставить
    0.82
    Activations Density 0.059%

    No Known Activations