© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 66822
    Prev
    Next
    INDEX
    Explanations

    config loaders or files

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

    This neuron responds to unusually long tokens (i.e. multi‐character subwords) such as lengthy identifiers, technical terms, or long words.

    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
     geene
    -1.14
     genoux
    -1.09
    綃
    -1.08
     prédé
    -1.05
     triom
    -1.04
    нутрен
    -1.02
     osal
    -1.02
     ennemi
    -1.01
     mortos
    -1.01
     boute
    -1.01
    POSITIVE LOGITS
    \\
    0.97
     fundamentales
    0.93
     {
    0.92
     integridad
    0.89
     re
    0.88
     "
    0.87
     マ
    0.87
     später
    0.85
    ERIE
    0.85
                       
    0.84
    Activations Density 0.119%

    No Known Activations