© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 93790
    Prev
    Next
    INDEX
    Explanations

    Song dynasty China

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

    The neuron is selectively activating on numeric tokens—especially year or date digits (often in parentheses) denoting reign periods or calendar years.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    Clark
    -0.83
     Telef
    -0.76
    etas
    -0.75
    Vogue
    -0.75
     отвер
    -0.74
     Turin
    -0.73
     blankets
    -0.72
     Bata
    -0.72
     teatral
    -0.72
     tapas
    -0.71
    POSITIVE LOGITS
     Song
    1.23
    宋
    1.19
     宋
    1.09
     song
    0.92
    Song
    0.91
    開封
    0.89
     Tang
    0.85
     officials
    0.80
    song
    0.80
     Northern
    0.79
    Activations Density 0.027%

    No Known Activations