© 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. 15988
    Prev
    Next
    INDEX
    Explanations

    why

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

    The neuron fires on standout descriptive content words—particularly adjectives or past‐participle verbs—that carry strong semantic weight (e.g. stylish, cute, nominated, light, reading, electrified).

    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
     पत्तियों
    0.89
    ból
    0.86
    ным
    0.85
    ных
    0.84
    ందని
    0.79
    larından
    0.79
     Allie
    0.77
    웁
    0.77
    되며
    0.77
    ными
    0.76
    POSITIVE LOGITS
     将
    0.81
     將
    0.76
    ﺃ
    0.75
     étrang
    0.74
     什么
    0.74
     本
    0.73
     問題
    0.72
    话
    0.71
     类型
    0.71
     précisément
    0.70
    Activations Density 0.000%

    No Known Activations