© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 27545
    Prev
    Next
    INDEX
    Explanations

    is a

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

    This neuron consistently fires on common, content-bearing function words and auxiliaries in technical or formal prose—e.g. “was,” “which,” “may,” “code,” “dealt”—i.e. moderately frequent non-stopwords that structure formal definitions or descriptions.

    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
    f
    -1.55
    a
    -1.46
    C
    -1.45
    N
    -1.39
    n
    -1.37
    an
    -1.34
    "
    -1.34
    e
    -1.34
    c
    -1.32
    t
    -1.31
    POSITIVE LOGITS
    pendentif
    1.28
    adaptador
    1.19
    ્યા
    1.17
    panasonic
    1.16
    〮
    1.16
    mujeres
    1.15
     Puoi
    1.13
    Ogni
    1.13
     mrt
    1.13
     enfrenta
    1.13
    Activations Density 0.294%

    No Known Activations