© 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
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 35016
    Prev
    Next
    INDEX
    Explanations

    world war

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

    The neuron is sensitive to capitalized words—particularly proper names and titles (e.g. game or book titles and author names).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    
    -0.90
    ίο
    -0.84
    聶
    -0.83
    יצוני
    -0.80
    ="'+
    -0.78
    Şi
    -0.75
     aktive
    -0.74
    ragon
    -0.74
     વિ
    -0.74
    Jangan
    -0.73
    POSITIVE LOGITS
     POD
    0.98
     TL
    0.92
    咦
    0.87
     GUR
    0.85
    vatore
    0.85
     butterflies
    0.84
     ReactDOM
    0.75
     acheter
    0.75
     Confederation
    0.75
     буты
    0.72
    Activations Density 0.025%

    No Known Activations