© 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. 22-GEMMASCOPE-RES-131K
    7. 130060
    Prev
    Next
    INDEX
    Explanations

    This neuron spotlights the start of new formal or mathematical statement sections—e.g. blank-line breaks, words like “Consider,” “Let,” “Now,” and the math‐mode “$” marker.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    questions and examples

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     gwiazd
    -0.96
    なので
    -0.96
     wouldn
    -0.96
    zarchiwizowane
    -0.95
    による
    -0.94
    ฬ
    -0.93
    但这
    -0.92
    が
    -0.91
    oramic
    -0.90
     الرئيسي
    -0.90
    POSITIVE LOGITS
    [])
    1.20
     energías
    1.18
     Tienes
    1.14
     your
    1.13
     Herzens
    1.09
    例如
    1.09
     például
    1.08
    bezpiecz
    1.07
     stalo
    1.05
    になったら
    1.05
    Activations Density 0.077%

    No Known Activations