© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 5621
    Prev
    Next
    INDEX
    Explanations

    The neuron detects or predicts patterns related to **foreign characters and words.**

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

    This neuron consistently lights up on the final content word just before a sentence or clause boundary.

    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
     څنګه
    0.81
    وں
    0.79
     Saltar
    0.79
     どれ
    0.79
     child
    0.79
     すぎ
    0.79
    ায়
    0.77
     項
    0.77
     bambino
    0.77
     flower
    0.76
    POSITIVE LOGITS
    сон
    0.77
    tson
    0.77
    maxlen
    0.76
    तः
    0.76
    Plate
    0.72
    වාස
    0.72
    wss
    0.71
    type
    0.71
    LAH
    0.71
    fragments
    0.70
    Activations Density 0.000%

    No Known Activations