© 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. 13197
    Prev
    Next
    INDEX
    Explanations

    unique words, success rate

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

    This neuron responds strongly to unusual or low‐frequency tokens—words or symbols that are rare in the input.

    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
     contradicted
    0.68
    ında
    0.67
    ͕
    0.67
     horrified
    0.67
    一组
    0.66
     nào
    0.64
    的数据
    0.63
    aturan
    0.63
     Những
    0.63
    seiten
    0.63
    POSITIVE LOGITS
    เบียน
    0.86
    ᱗
    0.79
    ០
    0.77
    また
    0.76
    ᱖
    0.76
    щё
    0.75
    ᱓
    0.74
    زاده
    0.73
    ৩৭
    0.73
    ῷ
    0.73
    Activations Density 0.000%

    No Known Activations