© 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. 34-GEMMASCOPE-RES-131K
    4. 118123
    Prev
    Next
    INDEX
    Explanations

    words starting with Kus- or Cus-

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

    The neuron responds to any occurrence of the letter sequence “cus” (i.e. tokens containing the substring “cus”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    Fran
    -0.77
     perspectiva
    -0.73
    ToInt
    -0.73
    ouds
    -0.67
    teos
    -0.67
    CB
    -0.66
    东京
    -0.66
     牙
    -0.66
    dente
    -0.66
     dépend
    -0.66
    POSITIVE LOGITS
     cus
    1.02
    Cus
    0.95
     Cus
    0.94
     Cushion
    0.88
    cus
    0.85
     Cousin
    0.84
     Cushi
    0.84
    Upstream
    0.79
    tgär
    0.79
    custo
    0.77
    Activations Density 0.015%

    No Known Activations