© 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. 104451
    Prev
    Next
    INDEX
    Explanations

    initials followed by last name

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

    The neuron fires on isolated uppercase single-letter tokens (often with following periods) that serve as initials in personal names or bibliographic citations.

    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
     其他
    -1.08
     also
    -1.06
     frecuente
    -1.02
    んでしょう
    -0.95
     déclaration
    -0.94
     similarly
    -0.93
    OREM
    -0.92
     daarbij
    -0.92
     habitations
    -0.91
     along
    -0.90
    POSITIVE LOGITS
    ariConfig
    1.09
     camisetas
    0.97
    Nichts
    0.95
    Neden
    0.94
     pokoje
    0.93
    Customers
    0.93
    ⃢
    0.93
    ادامه
    0.92
     cortado
    0.92
    اقتصاد
    0.91
    Activations Density 0.020%

    No Known Activations