© 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. 15644
    Prev
    Next
    INDEX
    Explanations

    cancel, remove, or hide

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

    The neuron fires on imperative action‐oriented method or function names (verbs like hide, detach, remove, disable, disconnect, etc.).

    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
     새로운
    -1.17
    させて頂
    -1.09
    シルエット
    -1.07
     informó
    -1.05
    either
    -1.05
    strengthen
    -1.04
    improved
    -1.02
    resulting
    -1.00
    letting
    -1.00
    Usos
    -1.00
    POSITIVE LOGITS
     all
    1.44
    了
    1.20
     or
    1.16
     EVERY
    1.11
     ALL
    1.10
     gruppo
    1.09
     any
    1.07
     each
    1.06
     the
    1.05
     currently
    1.05
    Activations Density 0.009%

    No Known Activations