© 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. 118799
    Prev
    Next
    INDEX
    Explanations

    oral sex or poor performance

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

    The neuron strongly activates on occurrences of the substring “suck” (in forms like suck, sucks, sucking, etc.).

    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
    媄
    -3.13
    ?”
    -2.95
    I
    -2.73
     have
    -2.55
     But
    -2.48
    苼
    -2.47
     years
    -2.44
    ING
    -2.39
    v
    -2.34
     were
    -2.33
    POSITIVE LOGITS
    絝
    2.61
    ୌ
    2.30
    䴀
    2.27
    the
    2.25
    ೢ
    2.19
    a
    2.14
    ⬈
    2.11
     padx
    2.11
    銠
    2.11
    kého
    2.06
    Activations Density 0.006%

    No Known Activations