© 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. 128904
    Prev
    Next
    INDEX
    Explanations

    The neuron selectively activates on numeric tokens—especially year or date numbers (e.g. “163”, “612”, “(XII)”, “728”, etc.)—in historical/wiki‐style text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    events, occurrences, and lists

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     only
    -1.11
    と呼ば
    -1.05
     you
    -1.04
     their
    -1.03
     now
    -1.02
    ToFit
    -1.02
    ıyoruz
    -1.00
     so
    -0.98
     formen
    -0.96
    塁
    -0.96
    POSITIVE LOGITS
     osserv
    1.36
    llac
    1.27
    adesivo
    1.25
     piacevole
    1.23
    Liefer
    1.21
     maggiore
    1.21
     consegna
    1.19
     marrone
    1.19
     ritorno
    1.18
     acontecimientos
    1.18
    Activations Density 0.024%

    No Known Activations