© 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. 22-GEMMASCOPE-RES-131K
    7. 38343
    Prev
    Next
    INDEX
    Explanations

    degree, improvement, cases

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

    This neuron detects sentence-initial discourse cues—words like “This,” “After,” “If,” “Just,” “In,” and “When” that signal transitions or introduce new clauses.

    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
     ahu
    -0.96
    istice
    -0.94
     huo
    -0.94
    approach
    -0.94
    пей
    -0.92
     Figura
    -0.91
     aborto
    -0.90
    eció
    -0.90
     kimi
    -0.89
    istica
    -0.88
    POSITIVE LOGITS
     who
    0.96
    婭
    0.93
     econômica
    0.88
     девочек
    0.87
    驳
    0.86
     Saxony
    0.85
     "{\"
    0.84
    reur
    0.83
     Frederick
    0.83
    落ちる
    0.82
    Activations Density 0.003%

    No Known Activations