© 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. 114148
    Prev
    Next
    INDEX
    Explanations

    Wasserstein distances

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

    The neuron is specially sensitive to tokens carrying embedded numeric values—especially those with “=%” format specifiers or other digit‐containing fragments.

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
     و
    -3.16
     \
    -2.94
     is
    -2.63
    ᯾
    -2.63
     '
    -2.42
     You
    -2.38
    ":
    -2.34
     letzte
    -2.33
     And
    -2.28
     và
    -2.28
    POSITIVE LOGITS
    ’，
    3.06
    ↵↵
    2.80
    ⡯
    2.75
    y
    2.69
    ーズン
    2.47
    StoryboardSegue
    2.44
    pecabe
    2.38
    彀
    2.30
    旎
    2.27
    踯
    2.27
    Activations Density 0.007%

    No Known Activations