© 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. 12991
    Prev
    Next
    INDEX
    Explanations

    wolf and shapeshifting

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

    The neuron fires on tokens relating to wolves (including “wolf,” “wolves,” and “were” as in “werewolves”).

    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
     വേണ്ട
    0.60
    टेट
    0.57
    стары
    0.56
    帚
    0.56
    ორცი
    0.55
    ᇁ
    0.55
     gerekir
    0.54
    arj
    0.53
    ECC
    0.53
    Список
    0.53
    POSITIVE LOGITS
     wolf
    1.29
    wolf
    1.16
    wolves
    1.11
     wolves
    1.10
    Wolf
    1.03
     Wolf
    1.02
    🐺
    0.98
     Wolves
    0.98
    狼
    0.97
    Wol
    0.93
    Activations Density 0.070%

    No Known Activations