© 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. 22-GEMMASCOPE-RES-131K
    7. 33849
    Prev
    Next
    INDEX
    Explanations

    names and specific entities

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

    The neuron strongly activates on tokens that begin new document sections or headings—that is, paragraph‐ or title‐initial words.

    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
     買
    -0.93
     usada
    -0.93
     tiế
    -0.89
    pineapple
    -0.88
     których
    -0.88
     odbior
    -0.86
     auttaa
    -0.85
    toddler
    -0.82
    rdd
    -0.82
    intéress
    -0.81
    POSITIVE LOGITS
    '}';
    1.02
    -//
    0.95
    __/
    0.91
    žia
    0.90
    nière
    0.90
    endforeach
    0.89
     フィギュア
    0.88
    ässä
    0.88
     Roca
    0.88
    ಅ
    0.87
    Activations Density 0.065%

    No Known Activations