© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 9534
    Prev
    Next
    INDEX
    Explanations

    surprise and shock

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

    This neuron detects words and phrases expressing strong emotional reactions or intensifiers (especially surprise or excitement).

    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_31_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
    थान
    0.54
     controller
    0.52
    螟
    0.52
     analyzed
    0.51
     UserController
    0.51
    controller
    0.50
    حديد
    0.50
    煩
    0.49
    iliter
    0.49
    시고
    0.48
    POSITIVE LOGITS
     surprise
    3.28
     surprised
    3.03
    surprise
    2.98
     surprises
    2.95
     sorpresa
    2.91
     shock
    2.86
     Surprise
    2.86
     shocked
    2.80
    Shock
    2.69
     sorpre
    2.69
    Activations Density 0.373%

    No Known Activations