© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 8637
    Prev
    Next
    INDEX
    Explanations

    cookies

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

    This neuron fires strongly on common legal-policy function words (e.g. will, use, or, the, a, is) characteristic of formal boilerplate statements.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    text from privacy policies and terms of service documents, particularly sections about data collection, cookies, and user rights.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    website privacy-policy and data-protection content, especially explanations of cookies, tracking/analytics, personal data handling, user consent, and legal compliance statements.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     vrchol
    0.71
     violently
    0.66
     đỉnh
    0.66
     vrch
    0.66
    ጅም
    0.64
     верши
    0.62
     вершины
    0.62
     Espí
    0.61
     calomel
    0.60
     furent
    0.59
    POSITIVE LOGITS
     cookies
    2.80
     cookie
    2.56
     Cookies
    2.56
    cookies
    2.53
    Cookies
    2.50
    cookie
    2.39
    Cookie
    2.36
     Cookie
    2.27
    🍪
    1.94
    COOKIE
    1.78
    Activations Density 0.350%

    No Known Activations