© 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. 32837
    Prev
    Next
    INDEX
    Explanations

    plot descriptions or introductions

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

    The neuron fires on document‐level metadata and structural markers—things like section headings, numeric scores, titles, and other formatting tokens rather than regular prose.

    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_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Negative Logits
     when
    -1.22
     even
    -1.02
    ᧐
    -0.92
    C
    -0.88
     …………
    -0.88
     當
    -0.88
     Even
    -0.87
     題
    -0.85
    例
    -0.84
    ")){
    -0.84
    POSITIVE LOGITS
    Spesifikasi
    1.14
    Đây
    1.11
     dibujado
    1.04
     vertes
    1.00
     Empieza
    0.99
    𓇼
    0.99
     interrom
    0.98
    两人的
    0.97
     skvě
    0.96
     vuelto
    0.96
    Activations Density 0.017%

    No Known Activations