© 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. 4516
    Prev
    Next
    INDEX
    Explanations

    code snippets with 'remove' or 'assume'

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

    This neuron fires on the final content word just before a sentence boundary (i.e. the last token before punctuation ending a sentence).

    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
    ла
    0.90
    га
    0.83
    кам
    0.83
    ographique
    0.77
    𝘭
    0.76
    亠
    0.75
    hner
    0.75
    𝒍
    0.74
    ны
    0.73
    рии
    0.73
    POSITIVE LOGITS
     wy
    0.78
     Into
    0.78
     Dish
    0.75
    ใน
    0.75
    AIDS
    0.74
     wad
    0.73
     ใน
    0.73
     HEK
    0.73
     Ин
    0.73
     disregard
    0.72
    Activations Density 0.000%

    No Known Activations