© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 72472
    Prev
    Next
    INDEX
    Explanations

    win record homology cycle triple

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

    The neuron detects highly specialized mathematical terminology or formal academic jargon—especially names of advanced constructs (e.g. “spectral triple,” “KK theory,” “Floer homology”).

    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
     can
    -1.39
     since
    -1.23
     podremos
    -1.17
     in
    -1.16
    ("
    -1.16
    Since
    -1.13
    ?!
    -1.11
     sodass
    -1.10
     můžete
    -1.10
     $
    -1.10
    POSITIVE LOGITS
    gibt
    1.38
    honestly
    1.34
     gruppi
    1.31
    believe
    1.28
    面對
    1.28
     saluti
    1.27
    letar
    1.25
    majority
    1.25
    somewhat
    1.23
    haben
    1.23
    Activations Density 0.015%

    No Known Activations