© 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. 22-GEMMASCOPE-RES-131K
    7. 25342
    Prev
    Next
    INDEX
    Explanations

    State or Slovakia

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

    The neuron is a “content‐word” detector that lights up on most non–function words—especially proper names, numbers, and other information‐bearing nouns and verbs—rather than on common articles or prepositions.

    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
     as
    -2.53
     epitom
    -2.38
     eight
    -2.36
     on
    -2.30
     in
    -2.30
     two
    -2.30
     at
    -2.27
     an
    -2.27
     one
    -2.27
     only
    -2.23
    POSITIVE LOGITS
     utiliser
    2.13
     Dinas
    2.09
     faciliter
    2.02
     créer
    1.98
     goofy
    1.98
     verhind
    1.96
    甗
    1.95
     ſen
    1.94
     alguns
    1.92
     CURRENT
    1.91
    Activations Density 0.134%

    No Known Activations