© 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. 26503
    Prev
    Next
    INDEX
    Explanations

    listing names and roles with commas and 'and'

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

    The neuron activates on numeric tokens (numbers and counts).

    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
     if
    -1.55
     many
    -1.52
     people
    -1.47
     while
    -1.38
     so
    -1.35
     some
    -1.34
     various
    -1.34
     for
    -1.33
     numerous
    -1.30
     time
    -1.29
    POSITIVE LOGITS
    originally
    1.27
    钡
    1.24
    couldn
    1.23
     who
    1.22
    ᡴ
    1.21
    quares
    1.20
     thuy
    1.19
     ſol
    1.19
    unfortunately
    1.16
     ANTI
    1.15
    Activations Density 0.038%

    No Known Activations