© 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. 8513
    Prev
    Next
    INDEX
    Explanations

    dialog or modal

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

    The neuron flags the “content” tokens in a passage—nouns (including names), numbers, and other information‐bearing words—while ignoring common function words.

    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
    𝐈
    0.83
    I
    0.79
    ̣ng
    0.74
    𝐟
    0.73
    中国
    0.72
     I
    0.71
    ̀ng
    0.71
    China
    0.70
    플
    0.70
    ชุด
    0.68
    POSITIVE LOGITS
    وتا
    0.91
     Corrosion
    0.85
    ısına
    0.81
     Cob
    0.80
     пути
    0.79
     Wett
    0.79
    اتها
    0.79
     अंधे
    0.77
     Sailor
    0.76
     wett
    0.75
    Activations Density 0.001%

    No Known Activations