© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 2684
    Prev
    Next
    INDEX
    Explanations

    ... at the end

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

    The neuron detects temporal expressions—words or phrases that refer to time, such as “yesterday,” “today,” “before,” specific times like “1 pm,” and duration indicators like “days.”

    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
     unless
    1.05
     &/
    1.03
    !!!!
    0.93
    !!!!!!
    0.91
     maupun
    0.90
     !!!!
    0.90
    !!!!!!!!
    0.89
     !!!!!
    0.89
     kecuali
    0.88
    !!!!!
    0.87
    POSITIVE LOGITS
     expecting
    0.83
    Expecting
    0.71
    觉得
    0.69
    すごく
    0.64
     intending
    0.63
    覺得
    0.62
    諗
    0.62
     hoping
    0.60
     pensando
    0.60
    Using
    0.58
    Activations Density 0.114%

    No Known Activations