© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 2378
    Prev
    Next
    INDEX
    Explanations

    ** headline formatting

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

    The neuron strongly activates on speaker‐attribution tokens and sentence boundaries—e.g. “said,” punctuation ending quotes, and paragraph breaks—marking reported speech attribution.

    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
    Reuters
    0.90
     Reuters
    0.80
    🛰
    0.75
    🚄
    0.74
     Bloomberg
    0.71
     বাংলাদেশের
    0.71
     internazionali
    0.71
     geopol
    0.70
    Bloomberg
    0.70
    ⏳
    0.68
    POSITIVE LOGITS
     Sandusky
    0.86
     Town
    0.77
     Oswego
    0.76
     County
    0.75
     Sgt
    0.74
    County
    0.71
     Editor
    0.71
     Danville
    0.70
     Hometown
    0.70
     Councilor
    0.70
    Activations Density 0.266%

    No Known Activations