© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 6973
    Prev
    Next
    INDEX
    Explanations

    punctuation and end tokens

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

    The neuron strongly activates on punctuation tokens—especially sentence‐ending marks (periods, commas, exclamation points, and closing quotes).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    '
    -3.25
    ↵↵
    -3.00
    0
    -2.78
    1
    -2.73
    很棒
    -2.39
    2
    -2.25
    ちなみに
    -2.09
     nineteen
    -2.03
    The
    -1.98
    ？！”
    -1.96
    POSITIVE LOGITS
     but
    2.70
     Also
    2.53
    
    2.36
    轳
    2.33
     According
    2.33
     Therefore
    2.25
     Aunque
    2.22
    をしている
    2.09
    taines
    2.09
     También
    2.09
    Activations Density 0.016%

    No Known Activations