© 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
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 99709
    Prev
    Next
    INDEX
    Explanations

    WC followed by "used"

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

    The neuron specifically detects the token “WC” (in any capitalization) when it appears as a standalone item or with adjoining digits—i.e. markers of word-count or numbered “WC” labels.

    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
    鲎
    -3.33
    咴
    -3.25
    ཋ
    -2.98
    骓
    -2.92
    蘩
    -2.75
    蹀
    -2.69
    縂
    -2.63
    杙
    -2.63
    嚒
    -2.59
    ൎ
    -2.58
    POSITIVE LOGITS
     dritten
    2.83
    or
    2.53
    里有
    2.48
     WC
    2.45
    id
    2.44
    co
    2.41
    o
    2.39
    WC
    2.38
    wc
    2.38
     
    2.34
    Activations Density 0.003%

    No Known Activations