© 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. 22-GEMMASCOPE-RES-131K
    4. 84084
    Prev
    Next
    INDEX
    Explanations

    abstract concepts and definitions

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

    This neuron lights up on the main “meta-discussion” terms of academic or analytic text—words naming core discussion points or concepts (e.g. “solution,” “question,” “view,” “methodological,” “intersection”).

    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
     and
    -1.89
     also
    -1.41
     will
    -1.37
     including
    -1.31
     has
    -1.28
     was
    -1.16
     are
    -1.16
     as
    -1.16
     it
    -1.15
     complete
    -1.15
    POSITIVE LOGITS
     that
    1.52
     avocat
    1.52
    謦
    1.50
     花朵
    1.49
     lampe
    1.45
     illustre
    1.45
     interdit
    1.44
     peint
    1.44
     panik
    1.42
     arabe
    1.41
    Activations Density 0.055%

    No Known Activations