© 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. 21076
    Prev
    Next
    INDEX
    Explanations

    foreign language and special characters

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

    This neuron activates on subword pieces of uncommon, specialized, or proper‐name terms (i.e. rare vocabulary segments).

    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
     and
    -2.72
     to
    -2.55
     on
    -2.39
     or
    -2.16
     at
    -2.02
    ).
    -2.02
     as
    -1.98
     all
    -1.93
     &
    -1.93
     you
    -1.89
    POSITIVE LOGITS
     pesso
    2.67
    횃
    2.33
    饪
    2.23
    俢
    2.23
    𖧵
    2.20
    苼
    2.14
    鸶
    2.14
     preparo
    2.14
    actéristi
    2.13
     coelho
    2.13
    Activations Density 0.891%

    No Known Activations