© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 13328
    Prev
    Next
    INDEX
    Explanations

    see, look, behold

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

    The neuron fires on attention-getting discourse markers—especially imperative “look/see/behold” style cues that directly address the reader.

    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
     zowel
    -1.14
     Talvez
    -1.10
    ⌗
    -1.06
     Então
    -1.02
     Hoje
    -0.99
     camiset
    -0.98
     Eintrag
    -0.97
    bardziej
    -0.96
     Entsche
    -0.95
     Bestimm
    -0.95
    POSITIVE LOGITS
     already
    1.27
    once
    1.04
     you
    1.02
    !
    1.00
     there
    0.98
     già
    0.94
    jeune
    0.94
    蕨
    0.93
    趁
    0.92
    albeit
    0.92
    Activations Density 0.017%

    No Known Activations