© 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. 10-GEMMASCOPE-RES-131K
    4. 11782
    Prev
    Next
    INDEX
    Explanations

    Boolean values and Director Michael

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

    The neuron fires on capitalized words (i.e. tokens beginning with an uppercase letter, like proper nouns and acronyms).

    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
     Both
    -1.80
     While
    -1.68
     лишь
    -1.66
     Since
    -1.56
     Those
    -1.47
    鋁
    -1.45
     Although
    -1.45
     enormously
    -1.43
     Firstly
    -1.39
     Though
    -1.37
    POSITIVE LOGITS
    occurrence
    1.27
     is
    1.21
    Soviet
    1.20
    蹰
    1.20
     keď
    1.19
     seleccionado
    1.18
    believe
    1.16
     hade
    1.16
    ときに
    1.15
    CDF
    1.14
    Activations Density 0.182%

    No Known Activations