© 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. 12858
    Prev
    Next
    INDEX
    Explanations

    field extensions like Q

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

    The neuron fires on tokens indicating division or fractional relationships—e.g. the slash “/” and the word “over.”

    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
     தாம
    0.66
     "*************
    0.60
     क्रेडिट
    0.59
     tendría
    0.58
     '**
    0.58
     podría
    0.57
     encontraron
    0.56
     antena
    0.55
     retras
    0.54
     posizione
    0.54
    POSITIVE LOGITS
    Aile
    0.59
     Highlight
    0.59
    ﻫ
    0.59
     Wolfgang
    0.58
    Highlight
    0.57
    Nowadays
    0.57
    ῇ
    0.56
    Oc
    0.55
     Shed
    0.55
    encio
    0.55
    Activations Density 0.000%

    No Known Activations