© 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. 128870
    Prev
    Next
    INDEX
    Explanations

    putative followed by technical terms

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

    The neuron strongly activates on longer, content‐heavy tokens—whether uncommon multi‐syllabic words (e.g. “putative,” “sectarian”) or capitalized proper names/terms—i.e. unusually long or technical tokens.

    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
    鸶
    -2.77
     que
    -2.64
     meas
    -2.45
    珰
    -2.42
     scrutin
    -2.30
    某个
    -2.28
     After
    -2.27
    ⬯
    -2.27
     một
    -2.25
     dejando
    -2.23
    POSITIVE LOGITS
    t
    3.03
    鉭
    2.61
    ⦅
    2.44
    錞
    2.41
    嫺
    2.39
     praktisch
    2.31
    晙
    2.25
    Ｊ
    2.23
     Analyzer
    2.20
     また
    2.20
    Activations Density 0.022%

    No Known Activations