© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B
    3. 31-GEMMASCOPE-2-RES-16K
    4. 12062
    Prev
    Next
    INDEX
    Explanations

    replace, return, remove

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

    The neuron fires on programming‐style syntax tokens—things like “#”, “(”, code keywords (Methods), and other punctuation or markup typical of code blocks.

    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
    :\
    0.64
    ').
    0.64
    :')
    0.61
     உள்ளனர்
    0.61
    AN
    0.60
    »).
    0.59
    。)
    0.57
    フィールド
    0.57
        
    0.56
    ');
    0.56
    POSITIVE LOGITS
     <--
    1.41
     noqa
    1.17
     ←
    1.08
     <-
    0.98
     Improves
    0.90
     Allows
    0.87
    <-
    0.87
     Tells
    0.84
    👈
    0.84
     Helps
    0.83
    Activations Density 0.013%

    No Known Activations