© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 44020
    Prev
    Next
    INDEX
    Explanations

    historical information and categories

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

    The neuron is sensitive to tokens that start new lines or sections—i.e. it fires strongly on first words or headings at the beginning of sentences or paragraphs.

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

    No Comments

    Negative Logits
     Ә
    -1.92
    匮
    -1.75
     who
    -1.58
    }`,
    -1.57
    𝛼
    -1.56
    әм
    -1.55
    dây
    -1.54
    TextEditing
    -1.53
    runOnUiThread
    -1.53
     Indonesia
    -1.53
    POSITIVE LOGITS
     of
    1.88
     правительства
    1.86
     современных
    1.68
     Istri
    1.65
     appartements
    1.59
    尔夫
    1.57
     Khart
    1.55
     logements
    1.43
     الحكومة
    1.42
     Theologie
    1.41
    Activations Density 0.648%

    No Known Activations