© 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. 65522
    Prev
    Next
    INDEX
    Explanations

    Q statements or introductions

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

    This neuron fires on tokens involved in accessing fields or properties—i.e. array‐index or object‐property lookups (the bracketed keys and “->name”/“->value” style accesses).

    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
     in
    -1.38
    祢
    -1.07
     with
    -1.05
    สิ่ง
    -1.02
     all
    -0.99
     for
    -0.97
     manera
    -0.97
    اته
    -0.94
    llten
    -0.93
    нена
    -0.90
    POSITIVE LOGITS
    ところに
    1.10
     atteindre
    1.05
     ánimo
    1.02
     ouvrir
    1.00
     dichi
    1.00
     visible
    1.00
     recevoir
    0.98
     lesquelles
    0.98
     FBref
    0.98
     benzina
    0.96
    Activations Density 0.002%

    No Known Activations