© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 35963
    Prev
    Next
    INDEX
    Explanations

    click here to

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

    The neuron identifies user‐action instructions—imperative UI commands (e.g. “Press Ctrl+C,” “click here,” “drag files”) prompting the reader to interact with the interface.

    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
     bekas
    -1.01
     appels
    -1.00
     flott
    -1.00
     hó
    -0.98
     falle
    -0.98
    Sabes
    -0.97
    previously
    -0.97
     exista
    -0.97
     kering
    -0.96
    announced
    -0.96
    POSITIVE LOGITS
     to
    2.53
     or
    1.35
     để
    1.10
     continue
    0.97
    เพื่อ
    0.96
     willkommen
    0.95
     nếu
    0.92
    ihnach
    0.91
     Innenstadt
    0.90
     after
    0.90
    Activations Density 0.007%

    No Known Activations