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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 45305
    Prev
    Next
    INDEX
    Explanations

    foreign words across languages

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

    The neuron is detecting question indicators—especially question‐mark punctuation and tag questions like “right.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    ↵↵
    -3.16
     l
    -2.22
    –
    -1.99
     For
    -1.91
     in
    -1.85
    </h2>
    -1.84
     â
    -1.76
     Time
    -1.75
     s
    -1.73
     gebraucht
    -1.73
    POSITIVE LOGITS
     minore
    2.08
     adatto
    1.99
     augus
    1.97
    笊
    1.96
     konus
    1.95
    婓
    1.94
    箜
    1.92
     anello
    1.88
     verhind
    1.86
    鯔
    1.84
    Activations Density 0.006%

    No Known Activations