© 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. 6557
    Prev
    Next
    INDEX
    Explanations

    Arthur

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

    The neuron is detecting proper‐name tokens—that is, fragments of named entities such as people, places, organizations, medications, or celestial bodies.

    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
    戕
    0.61
     특정
    0.58
     belirl
    0.57
     جناب
    0.57
     eminently
    0.55
    ेंडर
    0.55
     penchant
    0.53
    镌
    0.53
    ंडर
    0.53
     immersive
    0.52
    POSITIVE LOGITS
    =
    0.69
    は
    0.67
    と
    0.64
    는
    0.64
     walaupun
    0.64
    :
    0.64
    .
    0.63
    や
    0.63
    은
    0.62
    0.62
    Activations Density 0.001%

    No Known Activations