© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 25688
    Prev
    Next
    INDEX
    Explanations

    defining or describing states

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

    The neuron selectively activates on key topical nouns—often uncountable or abstract concepts (e.g. “Oxygen,” “Money,” “Cancer,” “Energy”) that serve as the main subject of a sentence or section.

    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
     from
    -1.48
     which
    -1.34
     some
    -1.31
     those
    -1.23
    tersebut
    -1.20
     with
    -1.19
     these
    -1.19
     Desember
    -1.15
     using
    -1.15
     that
    -1.13
    POSITIVE LOGITS
    是一种
    1.26
     anunció
    1.15
     裤
    1.13
    の基本
    1.11
    ruly
    1.10
    是一个
    1.09
     není
    1.07
     ontem
    1.06
     inoltre
    1.06
    cuits
    1.05
    Activations Density 0.136%

    No Known Activations