© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 7679
    Prev
    Next
    INDEX
    Explanations

    cock, coco, k, kakeya, kagame

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

    This neuron responds strongly to words made up of repeated syllables (reduplicated forms) such as “Kakao,” “Kakuna,” “Cocoon,” “Cocos,” or “cockatoo.”

    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
     মুখ্য
    0.56
    Zi
    0.56
    就被
    0.54
    HDI
    0.52
    tena
    0.51
    ミリ
    0.50
    Piper
    0.50
    钅
    0.49
     Zi
    0.49
    drift
    0.49
    POSITIVE LOGITS
     कक्क
    0.61
     scratch
    0.55
    꿩
    0.54
     putting
    0.52
    სა
    0.52
     sober
    0.52
     cak
    0.52
    तीय
    0.51
    एलबी
    0.51
     cracked
    0.50
    Activations Density 0.092%

    No Known Activations