© 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. 10-GEMMASCOPE-RES-131K
    4. 39105
    Prev
    Next
    INDEX
    Explanations

    Bureau of Census

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

    The neuron fires strongly on key domain-specific nouns and labels (like subroutine names, section headings, class/module names, and other technical or topical terms), essentially marking the main “content” words that carry the document’s structure and subject.

    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
     then
    -2.09
     into
    -1.89
     being
    -1.62
    ‘
    -1.47
    ly
    -1.41
     ‘
    -1.37
    man
    -1.36
     up
    -1.35
     You
    -1.33
     similar
    -1.33
    POSITIVE LOGITS
     mereka
    1.64
     presencial
    1.56
     sejak
    1.46
     podes
    1.35
    jedno
    1.35
    𓄹
    1.34
     肖像
    1.34
     surpre
    1.34
     bolí
    1.34
    dekor
    1.33
    Activations Density 0.127%

    No Known Activations