© 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. 6751
    Prev
    Next
    INDEX
    Explanations

    affected by, asked by, even change

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

    This neuron consistently lights up on the second token after each <start> boundary—that is, it detects the very first content word in every new segment.

    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
    ő
    0.78
    �
    0.74
    ג
    0.74
    agne
    0.73
    ⚫
    0.72
    Quando
    0.71
    Algun
    0.71
    0.70
    고
    0.68
    동안
    0.68
    POSITIVE LOGITS
     пункт
    0.87
    nerv
    0.80
     indigestion
    0.80
     convulsions
    0.79
    ivatives
    0.78
    ஜ்மஹ
    0.78
     thorium
    0.77
     degenerative
    0.75
     hepatocytes
    0.75
     irreparable
    0.75
    Activations Density 0.000%

    No Known Activations