© 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. 1824
    Prev
    Next
    INDEX
    Explanations

    followed by names

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

    This neuron flags high-level “summary‐worthy” tokens—key names, dates, numbers, and discourse markers that signal important events or turning points.

    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
    腺
    0.64
     উপকূল
    0.62
    nextDouble
    0.60
    anée
    0.59
     recreational
    0.57
    ূন্য
    0.57
     অতঃ
    0.56
     চিং
    0.56
    꽤
    0.55
    
    0.55
    POSITIVE LOGITS
     David
    0.98
     Priyanka
    0.96
     Ирина
    0.91
     Everyone
    0.90
     John
    0.86
     Joseph
    0.86
     Michael
    0.86
     Jokowi
    0.84
     Barack
    0.84
     Rihanna
    0.84
    Activations Density 0.507%

    No Known Activations