© 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. 22-GEMMASCOPE-RES-131K
    7. 15952
    Prev
    Next
    INDEX
    Explanations

    questions and affirmations

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

    The neuron activates on words in expressions of surprise, incredulity, or speaker self-correction/clarification (e.g. “did you say…?”, “you read that right,” “what…?”).

    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
    erkek
    -0.95
     MB
    -0.89
     everything
    -0.87
    🚐
    -0.85
    샀
    -0.85
    ządu
    -0.84
    这样
    -0.81
    這樣
    -0.81
    SpringRunner
    -0.81
     萬
    -0.81
    POSITIVE LOGITS
    ?
    1.09
     Schnitt
    0.90
    meiras
    0.90
     gaga
    0.87
     nói
    0.85
    聞き
    0.84
    もう一度
    0.84
     termasuk
    0.84
     oseb
    0.83
     called
    0.82
    Activations Density 0.011%

    No Known Activations