© 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. 28376
    Prev
    Next
    INDEX
    Explanations

    online products and services

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

    The neuron fires on plural, noun tokens that serve as section or category headings (e.g. “Products,” “Businesses,” “Boxes,” “Cars,” “Deals,” etc.).

    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
    ’
    -1.08
     what
    -1.05
     three
    -1.03
    被告
    -1.02
     unusual
    -1.02
     
    -1.01
     some
    -1.00
     an
    -0.99
     trai
    -0.99
     unbelievable
    -0.97
    POSITIVE LOGITS
    Royaume
    1.26
    leyebilirsiniz
    1.20
    を
    1.14
    もしか
    1.12
    Soorten
    1.09
    ные
    1.08
     могли
    1.07
    确实是
    1.05
    だからこそ
    1.05
     promesse
    1.05
    Activations Density 0.001%

    No Known Activations