© 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. 13980
    Prev
    Next
    INDEX
    Explanations

    variable assignments like a =

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

    This neuron specifically spikes on the first substantive word of each new document or snippet, marking the start of a 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
    is
    0.90
    ↵↵
    0.80
     :
    0.78
     Cod
    0.75
    ipsis
    0.75
     Legal
    0.74
     aes
    0.74
     Action
    0.72
    ively
    0.72
    чень
    0.71
    POSITIVE LOGITS
    ㌔
    1.03
    кою
    0.85
     ngồi
    0.81
     mối
    0.81
    étaient
    0.81
    ри
    0.80
    াইয়া
    0.80
     էին
    0.79
    ㍉
    0.79
    箜
    0.78
    Activations Density 0.000%

    No Known Activations