© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 72880
    Prev
    Next
    INDEX
    Explanations

    code comments and markers

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

    The neuron predominantly activates on standalone numeric literals (numbers) in the text.

    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
     while
    -1.70
     just
    -1.54
     where
    -1.47
     their
    -1.46
     relatively
    -1.41
     almost
    -1.40
     had
    -1.38
     more
    -1.35
     most
    -1.34
     despite
    -1.34
    POSITIVE LOGITS
     delincuentes
    1.93
     insanely
    1.47
     evaluaciones
    1.41
     Dinas
    1.38
     erhi
    1.33
    尨
    1.30
     nimmer
    1.28
     myös
    1.28
    does
    1.27
     marvelous
    1.26
    Activations Density 0.000%

    No Known Activations