© 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. 34-GEMMASCOPE-RES-131K
    7. 45669
    Prev
    Next
    INDEX
    Explanations

    food & drink

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

    The neuron strongly responds to the ampersand character (“&”) when it’s used in titles or category headings (e.g. “Food & Dining,” “Food & Beverage,” “Food & Wine”).

    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
     Agreements
    -0.82
    ショナル
    -0.81
    methoden
    -0.80
     ปี
    -0.80
    zulegen
    -0.79
     沖縄
    -0.77
    aremos
    -0.77
     ²
    -0.77
    fähigkeit
    -0.76
     TAC
    -0.76
    POSITIVE LOGITS
     drink
    2.34
     drinks
    2.31
     beverages
    2.08
     beverage
    2.00
     food
    1.52
    drink
    1.44
     Drink
    1.38
    drinks
    1.36
    Drinks
    1.29
    Drink
    1.27
    Activations Density 0.021%

    No Known Activations