© 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. 22-GEMMASCOPE-RES-131K
    4. 73710
    Prev
    Next
    INDEX
    Explanations

    library/sensor identifiers

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

    The neuron fires on unusually long individual tokens (e.g. long XML namespace URIs, extended whitespace indentations, or other very long string tokens).

    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
     their
    -1.59
    You
    -1.34
     at
    -1.22
     October
    -1.13
     which
    -1.12
     because
    -1.10
     by
    -1.09
    There
    -1.09
     September
    -1.03
     stated
    -1.03
    POSITIVE LOGITS
    yska
    1.11
     corriger
    1.11
     inaug
    1.09
     conster
    1.05
     compromis
    1.04
     incont
    1.02
    feri
    1.02
     quintessential
    1.00
     ストライプ
    0.98
     miſ
    0.98
    Activations Density 0.001%

    No Known Activations