© 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. 73231
    Prev
    Next
    INDEX
    Explanations

    welcome

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

    The neuron fires on quantifier or numeric noun phrases—spans headed by words like two, seven, over, all, those, a pair, etc., that introduce an amount or group.

    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
    ↵
    -1.94
    '
    -1.62
    <bos>
    -1.49
     autres
    -1.32
     أو
    -1.30
    或
    -1.30
     erreichbar
    -1.29
     otros
    -1.28
     estos
    -1.28
     menuju
    -1.24
    POSITIVE LOGITS
     from
    1.72
    ⽥
    1.52
    atique
    1.50
    escuela
    1.45
     into
    1.38
    鉺
    1.34
    strates
    1.34
     with
    1.30
    卣
    1.30
    passphrase
    1.28
    Activations Density 0.026%

    No Known Activations