© 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. 10-GEMMASCOPE-RES-131K
    7. 87150
    Prev
    Next
    INDEX
    Explanations

    surprise, horror, enjoyment, delight, detriment

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

    The neuron detects personal-emotion reaction phrases introduced by “to my/your…” (e.g. “to my surprise,” “to your horror,” “much to my delight”).

    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
     to
    -2.78
     variously
    -1.59
     now
    -1.55
     for
    -1.50
     such
    -1.49
     about
    -1.43
     said
    -1.43
     not
    -1.40
     either
    -1.36
     sicherzustellen
    -1.36
    POSITIVE LOGITS
    栃木
    1.67
    Penjelasan
    1.66
    }
    1.53
    Really
    1.48
    Какой
    1.47
     สอง
    1.47
     certains
    1.45
     parecia
    1.45
     éta
    1.45
    Another
    1.45
    Activations Density 0.008%

    No Known Activations