© 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. 31-GEMMASCOPE-2-RES-16K
    4. 9568
    Prev
    Next
    INDEX
    Explanations

    The neuron predominantly fires on common high-frequency function words and discourse markers (short stop-words like “for,” “that,” “be,” “all,” “can,” “will,” etc.).

    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_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
     Oriente
    0.57
     Jovi
    0.53
    愼
    0.51
     Jug
    0.50
     Ad
    0.49
     स्‍ट
    0.49
     Adm
    0.48
     Fat
    0.48
    नाची
    0.48
    裘
    0.47
    POSITIVE LOGITS
    for
    2.20
     foreach
    1.91
    foreach
    1.88
     for
    1.86
    forEach
    1.81
    ForEach
    1.71
    FOR
    1.67
     For
    1.63
    For
    1.62
     FOR
    1.58
    Activations Density 0.185%

    No Known Activations