© 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. 777
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on sentence-initial or clause-initial discourse markers and transition words (e.g. “But,” “However,” “So,” “Anyway,” “This,” etc.) that signal shifts, contrasts, or topic changes.

    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
     ،
    0.89
    ,
    0.86
    ,
    0.83
    ،
    0.83
    (),
    0.82
     ,
    0.79
     ,
    0.74
    -,
    0.71
    ,//
    0.69
    *,
    0.68
    POSITIVE LOGITS
     setelah
    1.07
     despite
    1.04
     nonostante
    1.03
     after
    0.97
     when
    0.97
     apesar
    0.95
     since
    0.94
     while
    0.93
     dopo
    0.93
     несмотря
    0.92
    Activations Density 1.005%

    No Known Activations