© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 14199
    Prev
    Next
    INDEX
    Explanations

    past findings and disclosures

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

    This neuron detects formal section headings or titles (e.g. “Description of the Prior Art,” “The present invention,” “Past attempts,” etc.) in technical or academic documents.

    oai_token-act-pair · o4-miniTriggered 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
    Not in Any Lists

    No Comments

    Negative Logits
     obuv
    -1.12
     nejs
    -1.09
    で見る
    -1.05
     シンプル
    -1.03
    tront
    -1.02
     Beide
    -1.01
    なんですが
    -0.99
     vandaag
    -0.99
     atentos
    -0.99
    ternos
    -0.96
    POSITIVE LOGITS
     about
    1.13
     on
    1.11
    4
    1.09
     and
    1.06
    5
    1.04
    2
    1.02
     of
    1.00
    8
    1.00
     in
    0.99
     only
    0.98
    Activations Density 0.115%

    No Known Activations