© 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. 12687
    Prev
    Next
    INDEX
    Explanations

    melanoma cells

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

    medical terms related to melanoma and Japanese language characters.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron strongly fires on the Cyrillic bigram “мо,” i.e. Russian words containing the substring “мо.”

    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
    If
    -3.03
    which
    -2.52
    或者
    -2.45
    what
    -2.39
     wichtiger
    -2.36
     They
    -2.33
    </em>
    -2.31
    ($
    -2.30
     quot
    -2.28
    し
    -2.27
    POSITIVE LOGITS
    嘖
    2.72
    𝄆
    2.67
    ⼈
    2.56
    眛
    2.47
     коллекции
    2.45
    ⚥
    2.39
    螻
    2.39
    穑
    2.34
    𓐍
    2.33
    癭
    2.33
    Activations Density 0.003%

    No Known Activations