© 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. 13423
    Prev
    Next
    INDEX
    Explanations

    presence or absence

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

    The neuron is picking out the first few content words at the start of each sentence or clause (i.e. sentence-initial 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_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
     to
    -2.63
     from
    -2.25
     in
    -2.25
     which
    -1.94
     on
    -1.69
     that
    -1.67
     allow
    -1.58
     for
    -1.57
     provide
    -1.50
     after
    -1.48
    POSITIVE LOGITS
    meden
    1.49
    ໆ
    1.48
    tomorrow
    1.41
     befre
    1.40
     geladeira
    1.38
     consacré
    1.37
    我们
    1.34
    jandra
    1.34
     misa
    1.33
     craindre
    1.32
    Activations Density 0.281%

    No Known Activations