© 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-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 45017
    Prev
    Next
    INDEX
    Explanations

    grant permission, Japanese saying, angel, casting

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

    The neuron fires on non-English verb stems or inflected verb forms (e.g. Japanese, Dutch, Hindi, etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     several
    -2.09
     some
    -2.03
     only
    -2.03
     all
    -1.95
     even
    -1.78
     any
    -1.69
     just
    -1.65
     many
    -1.63
     give
    -1.62
     make
    -1.60
    POSITIVE LOGITS
     sobot
    1.96
     okna
    1.84
    컹
    1.71
     trám
    1.70
     horrid
    1.70
     fotografie
    1.66
    𐄁
    1.65
     tomto
    1.64
    ANGEL
    1.63
     tohoto
    1.61
    Activations Density 0.004%

    No Known Activations