© 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. 14485
    Prev
    Next
    INDEX
    Explanations

    mixtape horseradish naloxone loaf tape

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

    The neuron strongly activates on the final chunk of longer, uncommon multi-syllable nouns (e.g. the “-inthe” in “Absinthe,” “-oxone” in “naloxone,” “-eradish” in “horseradish,” “-catraz” in “Alcatraz,” “-tape” in “mixtape,” 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
    翆
    -2.17
    
    -2.17
    汌
    -2.08
    헷
    -1.95
    L
    -1.92
    瘊
    -1.84
     utilizados
    -1.77
    dass
    -1.76
    玊
    -1.75
    re
    -1.74
    POSITIVE LOGITS
     is
    2.03
     Our
    2.02
    ."
    1.98
    ’
    1.91
     Originally
    1.91
     a
    1.70
    ulipas
    1.70
    ти
    1.68
    </strong>
    1.67
     We
    1.64
    Activations Density 0.018%

    No Known Activations