© 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
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 10225
    Prev
    Next
    INDEX
    Explanations

    This neuron fires strongly on the very last content word of a comment or sentence—i.e. it detects the final token at the end of a segment.

    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
    ية
    0.76
     หน่อย
    0.68
    Particip
    0.66
     siquiera
    0.64
     particolarmente
    0.64
    +$
    0.63
    RESS
    0.63
    ชื่อ
    0.63
    ᇁ
    0.62
    istema
    0.62
    POSITIVE LOGITS
    干部
    0.82
    ttu
    0.80
     Mew
    0.79
     Bobcats
    0.78
     BOOLE
    0.78
     MIF
    0.77
     күн
    0.77
     STRU
    0.77
    themem
    0.74
     Фи
    0.73
    Activations Density 0.000%

    No Known Activations