© 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. 34-GEMMASCOPE-RES-131K
    7. 2998
    Prev
    Next
    INDEX
    Explanations

    Fade in, Prime Minister, Dr

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

    The neuron detects screenplay or script formatting elements (e.g. “Fade in,” shot headings, time‐code numbers).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     Multiply
    -0.77
    FM
    -0.73
     multiply
    -0.72
     zwró
    -0.72
     frequently
    -0.70
    sista
    -0.70
    anao
    -0.69
    である
    -0.69
     αι
    -0.69
    kembangan
    -0.69
    POSITIVE LOGITS
     INT
    1.39
     OMITTED
    1.17
    INT
    1.09
     EXT
    1.05
     fade
    1.03
     ANGLE
    0.99
     dissolve
    0.96
    establishing
    0.96
    EXT
    0.94
     fades
    0.93
    Activations Density 0.019%

    No Known Activations