© 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. 10-GEMMASCOPE-RES-131K
    4. 29127
    Prev
    Next
    INDEX
    Explanations

    choose, select, drag, drop

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

    The neuron is firing on two‐character (CJK) compound words—i.e. Chinese/Japanese bigram terms.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    ar
    -2.05
     you
    -1.73
     becomes
    -1.62
     at
    -1.55
     trovo
    -1.51
     ของ
    -1.48
    شير
    -1.46
     because
    -1.46
     should
    -1.44
     there
    -1.43
    POSITIVE LOGITS
    珰
    1.95
    gunaan
    1.88
    螣
    1.66
    denk
    1.64
     unſ
    1.64
    keiten
    1.63
    ⑊
    1.62
     brug
    1.61
    
    1.56
    eding
    1.55
    Activations Density 0.004%

    No Known Activations