© 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-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 99333
    Prev
    Next
    INDEX
    Explanations

    names starting with K

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

    The neuron fires on three‐letter sequences of the form K–vowel–K (e.g. “Kek,” “Kak,” “Kik”), i.e. the substring K-V-K within a token.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    arote
    -0.94
     composto
    -0.73
     巻き
    -0.71
    getDatabase
    -0.71
     appelé
    -0.70
    СК
    -0.69
     بروز
    -0.69
     те
    -0.68
     EARL
    -0.68
    падает
    -0.67
    POSITIVE LOGITS
     precise
    0.79
     maž
    0.75
     continual
    0.74
    athus
    0.74
     truly
    0.71
     selva
    0.70
     visst
    0.70
     neutro
    0.70
     occasionally
    0.70
     berusia
    0.70
    Activations Density 0.021%

    No Known Activations