© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 68941
    Prev
    Next
    INDEX
    Explanations

    preventing or inhibiting actions

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

    This neuron fires on occurrences of the negation “won’t” (the “t” in “won’t”) and its associated verbs (as in “won’t give up” or “won’t shut its doors”).

    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
     rhino
    -0.98
    सि
    -0.86
    Approve
    -0.85
     établissements
    -0.82
     reinstatement
    -0.82
    ݂
    -0.81
     fave
    -0.81
    が多
    -0.80
     apparition
    -0.79
     vilka
    -0.79
    POSITIVE LOGITS
     restrictions
    1.09
     inhibitors
    1.01
     inhibitor
    0.84
    ＾
    0.84
     with
    0.84
     brakes
    0.80
     inhibition
    0.80
     inhibitory
    0.79
    ̑
    0.77
     deleteById
    0.77
    Activations Density 0.072%

    No Known Activations