© 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. 36491
    Prev
    Next
    INDEX
    Explanations

    abstract concepts and states

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

    This neuron fires on the passage’s key subject words—i.e. the main noun or proper‐name that defines what the paragraph or heading is about.

    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
    향
    -0.87
    而是
    -0.81
     Showing
    -0.81
    yu
    -0.79
    느
    -0.77
    ыл
    -0.76
     idiots
    -0.75
    
    -0.75
    '>"
    -0.75
    radio
    -0.75
    POSITIVE LOGITS
     firsthand
    1.08
     which
    1.01
     know
    1.01
     imagine
    0.99
     которая
    0.93
    putes
    0.92
     pierwszej
    0.92
     understand
    0.91
     olduğunu
    0.90
     그것
    0.89
    Activations Density 0.107%

    No Known Activations