© 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. 10-GEMMASCOPE-RES-131K
    7. 69721
    Prev
    Next
    INDEX
    Explanations

    preceded by parentheses or dashes

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

    The neuron fires on the little numeric metadata tags (e.g. the three‐digit story‐length or timing codes) that appear in the byline/header of newswire articles.

    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
     in
    -1.66
     who
    -1.48
     kteří
    -1.45
    </em>
    -1.38
     keep
    -1.38
     filho
    -1.38
    錨
    -1.24
     usually
    -1.19
     work
    -1.16
     kept
    -1.16
    POSITIVE LOGITS
    :
    1.60
    袛
    1.50
     uye
    1.48
    rrbracket
    1.34
     endast
    1.31
    を行い
    1.30
     cauſe
    1.27
    yrs
    1.27
     kangoo
    1.27
     ſtre
    1.23
    Activations Density 0.056%

    No Known Activations