© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 93106
    Prev
    Next
    INDEX
    Explanations

    arguments and opinions

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

    The neuron selectively fires on the opening “lede” lines or titles of article excerpts or posts, effectively marking the beginning of each new excerpt.

    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
    their
    -1.13
    $};
    -1.06
    ό
    -1.05
     chèvre
    -1.01
     immerhin
    -0.99
     pixabay
    -0.99
    herstellung
    -0.99
     Ministero
    -0.97
    蒸し
    -0.97
     pleaf
    -0.96
    POSITIVE LOGITS
     what
    1.40
     should
    1.28
     deberían
    1.09
     powin
    1.09
     Should
    1.05
     debería
    1.05
     ملع
    1.05
     πρέπει
    1.03
     America
    1.02
     society
    1.02
    Activations Density 0.156%

    No Known Activations