© 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. 22-GEMMASCOPE-RES-131K
    7. 127685
    Prev
    Next
    INDEX
    Explanations

    The neuron identifies section‐ or heading‐style tokens—especially those labeling numbered sections or titles (e.g. “Level 1:”, “Group C”, standalone numbers that head subsections).

    oai_token-act-pair · o4-mini
    Triggered by @jyhe0408

    list headings

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Negative Logits
    ated
    -0.85
    kuuta
    -0.85
     μό
    -0.83
     stronę
    -0.79
     cible
    -0.79
    Руко
    -0.79
    Digite
    -0.76
    План
    -0.76
     was
    -0.75
     kem
    -0.75
    POSITIVE LOGITS
     each
    1.19
    Below
    1.08
     herewith
    1.00
     below
    0.99
    taux
    0.98
     here
    0.98
     Below
    0.97
    第一天
    0.96
    各
    0.94
     clockwise
    0.93
    Activations Density 0.324%

    No Known Activations