© 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. 10788
    Prev
    Next
    INDEX
    Explanations

    proper nouns and titles

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

    The neuron mainly fires on capitalized tokens—proper names, places, dates, brands, and other named‐entity 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_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
     our
    -2.31
     gives
    -2.03
     as
    -1.86
     gave
    -1.77
     give
    -1.77
     putting
    -1.76
     brings
    -1.73
     provides
    -1.72
     makes
    -1.69
     exquisitely
    -1.68
    POSITIVE LOGITS
    
    1.74
    訁
    1.73
     vijf
    1.61
     Dinas
    1.61
    commonly
    1.55
    originally
    1.53
     ส์
    1.48
     ۱۰
    1.48
     ۱۲
    1.46
     praticamente
    1.45
    Activations Density 0.074%

    No Known Activations