© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 74939
    Prev
    Next
    INDEX
    Explanations

    sequence beginnings

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

    The neuron fires strongly on sentence-boundary markers and the words immediately following them, effectively spotting the beginnings of sentences.

    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
     and
    -1.10
    ’?
    -1.00
    jména
    -0.94
     ?>/
    -0.94
    fficiency
    -0.91
    ficial
    -0.90
    $$$
    -0.87
     then
    -0.85
     still
    -0.85
     Influenza
    -0.84
    POSITIVE LOGITS
     initially
    1.18
     previously
    1.16
    気に入り
    1.10
    これを
    1.07
     inicialmente
    1.05
     originally
    1.02
     рестора
    1.00
    cellor
    0.93
     juez
    0.93
    ykjav
    0.91
    Activations Density 0.184%

    No Known Activations