© 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. 9038
    Prev
    Next
    INDEX
    Explanations

    dash punctuation

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

    The neuron strongly activates on dash‐like separators (single, double, en- or em-dashes) used to mark section or clause breaks in the text.

    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
     gives
    -2.52
     or
    -2.52
     will
    -2.31
     allows
    -2.28
     creates
    -2.25
     then
    -2.25
     seems
    -2.22
     becomes
    -2.20
     would
    -2.19
     only
    -2.19
    POSITIVE LOGITS
    
    2.06
    犴
    1.97
     Dinas
    1.92
    嫑
    1.91
    ᵴ
    1.80
     rése
    1.77
    艄
    1.76
     ſes
    1.74
    despite
    1.74
     indiquer
    1.73
    Activations Density 0.102%

    No Known Activations