© 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. 31-GEMMASCOPE-2-RES-16K
    4. 10643
    Prev
    Next
    INDEX
    Explanations

    expansion and contraction

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

    The neuron flags timestamp tokens in edit footers (e.g. the “HH:MM AM/PM” time strings).

    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
     因
    0.62
     Erinner
    0.59
     馬
    0.59
     Pferde
    0.57
    લિ
    0.57
    ዪ
    0.55
     echoing
    0.55
     RUDDER
    0.55
     马
    0.55
    马
    0.55
    POSITIVE LOGITS
    expansion
    0.63
    gren
    0.61
     expansion
    0.59
    ↵
    0.59
    Expansion
    0.59
     dil
    0.58
    uju
    0.57
     Expansion
    0.56
     expans
    0.55
     gren
    0.55
    Activations Density 0.000%

    No Known Activations