© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 72564
    Prev
    Next
    INDEX
    Explanations

    Thai language and culture

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

    This neuron fires on words ending in a two‐vowel sequence (diphthong), e.g. tokens ending in “ai,” “ay,” or “ie.”

    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
     eleições
    -2.80
     knows
    -2.52
    各样的
    -2.50
     are
    -2.48
    地板
    -2.47
     Bedürfn
    -2.44
     تعدى
    -2.42
    廡
    -2.31
    全都
    -2.30
    </h3>
    -2.25
    POSITIVE LOGITS
    \
    2.92
    U
    2.34
    N
    2.28
    Several
    2.27
    A
    2.25
     setelan
    2.20
    のお話
    2.17
    Using
    2.17
     
    2.16
    I
    2.16
    Activations Density 0.004%

    No Known Activations