© 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. 130972
    Prev
    Next
    INDEX
    Explanations

    pretend or pretense

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

    This neuron detects occurrences of the word “pretend” (and its morphological variants) 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_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
    ꦢ
    -2.47
    }$
    -2.44
     fratelli
    -2.41
    replyreport
    -2.38
    籏
    -2.31
     když
    -2.30
    n
    -2.25
     この
    -2.25
    شنایی
    -2.23
    跄
    -2.17
    POSITIVE LOGITS
    </h3>
    2.77
    淠
    2.61
    .
    2.59
    的道
    2.30
    を取る
    2.25
    随即
    2.11
    
    2.09
     Thanks
    2.05
     της
    2.03
    ológicos
    2.02
    Activations Density 0.002%

    No Known Activations