© 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. 12800
    Prev
    Next
    INDEX
    Explanations

    periods and centuries

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

    The neuron selectively fires on numeric tokens that mark ordered events or positions (e.g. innings, rounds, minutes) in sports reporting.

    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
     halfway
    0.64
     WAR
    0.60
     борь
    0.59
     pacing
    0.59
     midway
    0.58
     fight
    0.57
     잘
    0.56
     あっ
    0.56
     halftime
    0.55
     NEAR
    0.54
    POSITIVE LOGITS
     ninth
    0.99
     seventh
    0.90
     eighth
    0.90
    eighth
    0.84
     sixth
    0.81
     Eighth
    0.80
    seventh
    0.80
     Ninth
    0.79
     Seventh
    0.78
    第九
    0.77
    Activations Density 0.078%

    No Known Activations