© 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. 38537
    Prev
    Next
    INDEX
    Explanations

    Pearl Harbor and World Wars

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

    The neuron strongly activates on numeric tokens—especially multi‐digit years and date components.

    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
     dinosaurio
    -0.85
     fugitive
    -0.79
    atoire
    -0.77
     susi
    -0.77
    gento
    -0.77
     fluctuations
    -0.77
     Helicopter
    -0.76
     ķ
    -0.75
     láser
    -0.74
    صفات
    -0.73
    POSITIVE LOGITS
     Pearl
    2.00
    Pearl
    1.65
     Oahu
    1.27
     pearl
    1.19
    pearl
    1.15
     PEAR
    1.10
    珍珠
    1.08
     Hawaii
    0.98
     December
    0.96
     FDR
    0.94
    Activations Density 0.014%

    No Known Activations