© 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. 33714
    Prev
    Next
    INDEX
    Explanations

    multilingual technical and foreign words

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

    The neuron activates on specialized jargon and proper‐name tokens—particularly technical acronyms, chemical or medical terms, product names, and other domain-specific keywords.

    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
     verschillende
    -1.06
     sommige
    -1.05
    nachron
    -1.02
     algumas
    -1.00
     alguns
    -0.96
     некоторые
    -0.95
     lessened
    -0.94
    IDAS
    -0.92
     bazı
    -0.91
     ウール
    -0.90
    POSITIVE LOGITS
     of
    1.29
    了一会儿
    0.89
     nevera
    0.89
    ⅝
    0.88
     statt
    0.88
     lugar
    0.88
    ChildScrollView
    0.87
     ministerio
    0.87
     кнопку
    0.86
    有意思
    0.86
    Activations Density 0.014%

    No Known Activations