© 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. 31-GEMMASCOPE-2-RES-16K
    7. 386
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on tokens that are capitalized mid‐sentence—that is, on the initial pieces of proper names or other internally uppercase words.

    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
     readonly
    0.46
     rech
    0.42
    0.41
     previs
    0.39
     TODO
    0.39
    <0x0D>
    0.39
     tetr
    0.38
     int
    0.38
    <unused62>
    0.37
     multidis
    0.37
    POSITIVE LOGITS
    ri
    0.52
    é
    0.51
    ney
    0.50
    న్
    0.49
    ifornia
    0.48
    én
    0.48
    idity
    0.48
    aiian
    0.48
    ric
    0.47
    ás
    0.47
    Activations Density 3.185%

    No Known Activations