© 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-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 913
    Prev
    Next
    INDEX
    Explanations

    The neuron is acting like a “surprisal” detector: it spikes on individual words that are especially rare or unexpected in context.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    рта
    0.92
    ypen
    0.80
    ția
    0.79
     શું
    0.78
     threx
    0.78
    夻
    0.78
     Nasıl
    0.77
     Mình
    0.77
    िरिक्स
    0.77
     Nhân
    0.76
    POSITIVE LOGITS
    ’
    1.16
    '
    0.84
    ;
    0.73
     pumpkin
    0.71
    ’;
    0.71
    мы
    0.70
     الطرق
    0.70
     Widers
    0.70
     string
    0.70
    ,
    0.68
    Activations Density 0.000%

    No Known Activations