© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 77050
    Prev
    Next
    INDEX
    Explanations

    offensive language and content

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

    This neuron activates on tokens related to abusive or disallowed content—things like “abusive,” “derogatory language,” “vulgar,” “hate speech,” “aggressive behavior,” or similarly negative/moderation‐triggering 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_34/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
    چی
    -0.77
    RIX
    -0.75
     мати
    -0.75
    ترت
    -0.73
     работы
    -0.73
     kebaikan
    -0.71
     upvote
    -0.71
     wikihow
    -0.71
     cerveau
    -0.71
    YourGuide
    -0.71
    POSITIVE LOGITS
     language
    5.13
    language
    3.38
     Language
    3.16
    Language
    3.14
     vocabulary
    3.06
     lenguaje
    3.03
     terminology
    2.92
     linguagem
    2.78
     LANGUAGE
    2.78
     wording
    2.72
    Activations Density 0.065%

    No Known Activations