© 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. 214
    Prev
    Next
    INDEX
    Explanations

    This neuron lights up on the punctuation and phrase boundaries that mark digressions or parenthetical asides—especially em-dashes, ellipses, and the words immediately before them.

    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
    、
    0.55
    ،
    0.49
    ，
    0.46
     ،
    0.40
     、
    0.40
    ,
    0.39
    (),
    0.39
            
    0.38
    의
    0.37
    )،
    0.37
    POSITIVE LOGITS
     but
    1.02
    but
    0.96
     mutta
    0.89
     लेकिन
    0.88
     pero
    0.86
     ولكن
    0.86
     אבל
    0.86
     لكن
    0.80
     nhưng
    0.79
     ஆனால்
    0.77
    Activations Density 1.563%

    No Known Activations