© 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. 67193
    Prev
    Next
    INDEX
    Explanations

    click and browse buttons

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

    The neuron fires on GUI interface action words and button labels (e.g. “Click,” “Browse,” “Button,” “Select”) indicating it’s detecting references to interactive button commands.

    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
    𓆏
    -1.16
     tych
    -1.13
    acons
    -1.01
    several
    -0.98
    看著
    -0.98
     órgãos
    -0.97
    those
    -0.97
     these
    -0.96
     strikingly
    -0.96
    gives
    -0.94
    POSITIVE LOGITS
     burgeoning
    1.13
     at
    1.13
     JUST
    1.07
     Getting
    1.01
     As
    0.99
     helps
    0.97
     then
    0.97
     Not
    0.96
     longstanding
    0.96
     endearing
    0.96
    Activations Density 0.005%

    No Known Activations