© 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. 10-GEMMASCOPE-RES-131K
    7. 74952
    Prev
    Next
    INDEX
    Explanations

    overlay, Field, down, engineer, clamp, meta, induction, brig

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

    The neuron responds to low‐frequency or unusually specific content words—things like rare technical terms, proper names, or other uncommon vocabulary.

    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
    雱
    -1.73
     rác
    -1.63
     nhưng
    -1.54
     手順
    -1.54
     permitir
    -1.52
     FRANC
    -1.49
     fizeram
    -1.48
     überprü
    -1.46
    傒
    -1.45
     garantías
    -1.43
    POSITIVE LOGITS
     not
    1.98
     no
    1.51
    s
    1.50
     when
    1.48
    us
    1.42
     Arxivat
    1.41
     on
    1.40
     et
    1.37
     in
    1.34
     my
    1.34
    Activations Density 2.658%

    No Known Activations