© 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. 22-GEMMASCOPE-RES-131K
    7. 64792
    Prev
    Next
    INDEX
    Explanations

    institution or location information

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

    This neuron fires on isolated uppercase letters or digit‐only tokens (e.g. single capitals and numbers), as typically found in section headings or metadata labels.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     have
    -1.38
     their
    -1.30
     destina
    -1.23
     any
    -1.22
    LEBR
    -1.15
     anything
    -1.13
     zaragoza
    -1.10
     deriva
    -1.09
     see
    -1.09
     such
    -1.05
    POSITIVE LOGITS
     ориги
    1.13
    ルド
    1.11
     русской
    1.08
     deixa
    1.07
    astă
    1.06
    empleado
    1.05
    urther
    1.03
    痤疮
    1.03
     partager
    1.02
    ネン
    1.01
    Activations Density 0.001%

    No Known Activations