© 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. 13520
    Prev
    Next
    INDEX
    Explanations

    sudden forceful actions

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

    The neuron is primarily detecting high-intensity physical action verbs and related terms—especially sudden, violent or forceful acts (e.g. “fired,” “shot,” “slams,” “ambushed,” “pump,” “cock”).

    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
     also
    -1.84
     frågan
    -1.07
     necessary
    -1.05
     especially
    -1.00
     here
    -0.99
     \&
    -0.97
     też
    -0.95
     particularly
    -0.94
     Also
    -0.94
     you
    -0.93
    POSITIVE LOGITS
    いきなり
    1.12
    dler
    1.07
     arrep
    1.05
    !
    1.05
     البرنامج
    1.01
    intos
    1.00
     المعلومات
    1.00
     getNome
    0.99
    仕上がり
    0.98
    nemone
    0.97
    Activations Density 0.011%

    No Known Activations