© 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. 10-GEMMASCOPE-RES-131K
    7. 9029
    Prev
    Next
    INDEX
    Explanations

    technical compound terms

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

    The neuron fires on programming‐style identifiers and type names—especially capitalized tokens like class or type names (e.g. “Type”, “NET”, “InlineData”) found in source code.

    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
    认同
    -1.70
    atized
    -1.67
    但在
    -1.58
    躙
    -1.53
    cartão
    -1.52
    焄
    -1.52
     деца
    -1.50
    Fácil
    -1.49
    COMO
    -1.46
    それは
    -1.45
    POSITIVE LOGITS
     your
    2.84
     a
    2.81
     his
    2.08
     another
    1.95
     such
    1.91
     an
    1.88
    "
    1.85
     our
    1.80
     this
    1.66
     again
    1.57
    Activations Density 5.673%

    No Known Activations