© 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. 31-GEMMASCOPE-2-RES-16K
    7. 5709
    Prev
    Next
    INDEX
    Explanations

    state or existence in Korean

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

    The neuron activates on possessive forms, i.e. tokens ending with an apostrophe-s.

    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
    garbage
    0.84
    垃圾
    0.83
     garbage
    0.81
    trash
    0.79
     trash
    0.78
    Trash
    0.78
    ゴミ
    0.76
     sampah
    0.74
    Waste
    0.68
    discard
    0.65
    POSITIVE LOGITS
     있는데
    1.19
     있는데요
    1.09
    었는데
    1.00
     하는데
    0.97
    는데
    0.97
    는데요
    0.95
    했는데
    0.95
    인데
    0.93
    인데요
    0.85
    하는데
    0.84
    Activations Density 0.001%

    No Known Activations