© 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. 98168
    Prev
    Next
    INDEX
    Explanations

    statement or charter and related nouns

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

    The neuron strongly picks out standalone, attention‐grabbing content nouns (e.g. “split,” “statement,” “charter,” “index,” “test,” “miracle,” “mystery”) that function as key thematic or headline words.

    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
    それも
    -1.49
    спомина
    -1.47
    8
    -1.45
    などを
    -1.40
     olympiques
    -1.40
    apparently
    -1.39
    Книга
    -1.39
    そこで
    -1.39
    もしか
    -1.38
    
    -1.38
    POSITIVE LOGITS
     of
    2.05
     our
    1.95
    q
    1.44
     miesięcy
    1.42
    人也
    1.41
    之地
    1.40
    williams
    1.38
     jakość
    1.38
    umpulan
    1.37
    🎊
    1.31
    Activations Density 0.089%

    No Known Activations