© 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-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 11809
    Prev
    Next
    INDEX
    Explanations

    user actions and interface elements

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

    The neuron is detecting words associated with user‐generated feedback or commentary (e.g., comments, posts, reviews, insults, slander, opinions).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_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
    OrDefault
    0.59
     kwa
    0.58
    ized
    0.57
    cache
    0.57
     kinetics
    0.56
    IZED
    0.55
     _;
    0.55
    cached
    0.54
    _{
    0.53
    Cache
    0.53
    POSITIVE LOGITS
    єднання
    0.73
    画面
    0.72
     кнопку
    0.72
    按钮
    0.71
    教室
    0.71
    ボタン
    0.71
    手続き
    0.70
    สนุก
    0.70
    を行う
    0.70
     textbox
    0.70
    Activations Density 0.472%

    No Known Activations