© 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. 15100
    Prev
    Next
    INDEX
    Explanations

    numbers, comments, starting phrases

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

    The neuron predominantly fires on numeric tokens—years, dates, counts, and other digit-containing items—wherever they appear in the text.

    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
     our
    -1.60
     we
    -1.29
    デザート
    -1.02
     قبل
    -1.01
     roślin
    -1.00
    pellier
    -0.99
     my
    -0.99
     десер
    -0.98
    Can
    -0.96
    bool
    -0.96
    POSITIVE LOGITS
     दोनों
    1.41
    ksikon
    1.30
    ṙ
    1.27
     دشمن
    1.18
    zahn
    1.18
     Beide
    1.15
     Janeiro
    1.15
    发布时间
    1.11
    உ
    1.10
    וּ
    1.09
    Activations Density 0.003%

    No Known Activations