© 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. 10-GEMMASCOPE-RES-131K
    7. 20813
    Prev
    Next
    INDEX
    Explanations

    many studies

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

    This neuron is tuned to the technical/research‐paper style register—spotting common scientific-article words (e.g. “compounds,” “known,” “exist,” “studies,” “have reported,” etc.) that occur in academic or formal research writing.

    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
     two
    -2.30
    了一個
    -1.85
     one
    -1.80
     all
    -1.79
     most
    -1.77
     three
    -1.72
     some
    -1.66
     sometimes
    -1.56
     first
    -1.52
    了一个
    -1.52
    POSITIVE LOGITS
     olika
    1.96
     įva
    1.66
     gerekiyor
    1.50
     farklı
    1.49
     múltip
    1.48
     INCLUDING
    1.48
     различ
    1.46
     различных
    1.45
     forskellige
    1.44
     نیز
    1.42
    Activations Density 0.820%

    No Known Activations