© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 9874
    Prev
    Next
    INDEX
    Explanations

    other things

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

    This neuron specializes in picking up discourse‐level connectors—words like “and,” “but,” “however,” “until,” “at least,” etc., that signal transitions or coordination in the text.

    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
    ות
    1.34
    没有
    1.32
     ώστε
    1.25
    ع
    1.23
    ج
    1.22
    נ
    1.22
    你
    1.22
    もちろん
    1.19
    스
    1.18
    ने
    1.17
    POSITIVE LOGITS
    romeda
    1.68
     sebagainya
    1.54
     whatnot
    1.45
    rogens
    1.26
    wiches
    1.24
    irons
    1.16
     Subsidi
    1.14
    ंगाबाद
    1.14
     seterusnya
    1.13
     pensez
    1.12
    Activations Density 15.528%

    No Known Activations