© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 85884
    Prev
    Next
    INDEX
    Explanations

    end-of-quote punctuation

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

    The neuron selectively activates on semantically rich “topic” words—especially standalone content nouns or key terms (often in quotes), such as air, bound, bit, mass, dream, etc.

    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
     one
    -2.00
     on
    -1.98
     using
    -1.88
     focusing
    -1.64
     which
    -1.61
     efficient
    -1.58
     we
    -1.57
    .
    -1.57
     increases
    -1.55
    Following
    -1.55
    POSITIVE LOGITS
     différents
    1.78
     THOUS
    1.71
     SEPAR
    1.66
     intertw
    1.65
    
    1.64
     détect
    1.63
     ส์
    1.63
    
    1.62
     SPECIFIC
    1.62
     SEVER
    1.61
    Activations Density 0.309%

    No Known Activations