© 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. 13416
    Prev
    Next
    INDEX
    Explanations

    foreign words and names

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

    The neuron strongly activates on non-English word fragments or tokens (e.g. German or Japanese), i.e. foreign-language text.

    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
    অত
    0.73
    BadRequest
    0.73
    Bigg
    0.70
    Introduceti
    0.68
    अपने
    0.67
    Nya
    0.67
    seits
    0.66
    ⤷
    0.66
     klasik
    0.66
     rezat
    0.65
    POSITIVE LOGITS
     家
    0.68
     世
    0.65
     연
    0.63
     ブ
    0.60
    )
    0.59
     फ़
    0.59
     電
    0.59
    筫
    0.59
     แ
    0.57
     학
    0.57
    Activations Density 0.150%

    No Known Activations