© 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. 15180
    Prev
    Next
    INDEX
    Explanations

    sports scores and statistics

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

    The neuron mainly activates on numeric tokens—especially numbers representing statistical values like points, yards, innings, etc.

    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
    -1.45
     at
    -1.43
     on
    -1.43
     after
    -1.42
     their
    -1.42
     because
    -1.36
     which
    -1.29
     it
    -1.28
     using
    -1.25
     to
    -1.23
    POSITIVE LOGITS
     tandis
    1.22
     zoude
    1.15
     interag
    1.11
    𓆏
    1.11
    
    1.07
     accompag
    1.06
     chacune
    1.05
     ホルダー
    1.05
     respecte
    1.05
     المعرف
    1.04
    Activations Density 0.010%

    No Known Activations