© 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. 22-GEMMASCOPE-RES-131K
    7. 38946
    Prev
    Next
    INDEX
    Explanations

    arithmetic and algebraic expressions

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

    The neuron fires on numeric tokens—especially integers and digits—when they appear as part of arithmetic calculations or intermediate computation results.

    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
    rozum
    -1.15
    kennen
    -1.13
    
    -0.97
    gdx
    -0.95
    czego
    -0.95
     ihm
    -0.95
     mortos
    -0.95
    تاح
    -0.95
     subscribe
    -0.94
    rufen
    -0.93
    POSITIVE LOGITS
     different
    1.17
     poteva
    1.02
     funktionieren
    1.02
     nuevamente
    1.01
    ὺς
    0.98
     doesn
    0.98
     again
    0.97
     drehen
    0.96
     but
    0.96
     nochmals
    0.96
    Activations Density 0.004%

    No Known Activations