© 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. 31-GEMMASCOPE-2-RES-16K
    7. 6031
    Prev
    Next
    INDEX
    Explanations

    verbs of saying followed by quotes

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

    The neuron detects the reporting‐speech attribution verb “said” (i.e. speaker attributions).

    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
     '[
    0.73
     [],
    0.72
     "..
    0.72
     stating
    0.70
     “[
    0.70
     diciendo
    0.68
     “.
    0.68
     ([
    0.66
     "...
    0.66
     "-",
    0.66
    POSITIVE LOGITS
     you
    0.66
     Usually
    0.64
     आमतौर
    0.63
     Обычно
    0.62
     সাধারণত
    0.61
    もちろん
    0.61
    まぁ
    0.60
     Поскольку
    0.60
     Расійскай
    0.60
     usually
    0.60
    Activations Density 0.045%

    No Known Activations