© 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. 1784
    Prev
    Next
    INDEX
    Explanations

    This neuron detects technical jargon and acronyms—particularly file-format names and other domain‐specific terms (e.g. “wav,” “mp3,” “WMA,” “DLNA,” “overlay,” “infrared,” “ultrasonic,” etc.).

    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
    avao
    0.80
    んにちは
    0.80
    ড়িয়ে
    0.77
    ᧐
    0.75
    nhs
    0.72
    ными
    0.71
    ॉक
    0.71
    urther
    0.70
    吖
    0.70
    ối
    0.70
    POSITIVE LOGITS
     ${
    0.72
     v
    0.70
     \
    0.69
     w
    0.63
     $[
    0.63
     [
    0.63
     k
    0.62
     p
    0.62
     (
    0.61
     proper
    0.60
    Activations Density 0.844%

    No Known Activations