Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    running code from terminal

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     problems
    0.90
     communities
    0.90
     fantastic
    0.86
     incredible
    0.85
     lawsuits
    0.82
     modes
    0.81
     difficult
    0.80
     difficulties
    0.78
     countries
    0.78
     communautés
    0.77
    POSITIVE LOGITS
     lounge
    1.68
     comedor
    1.60
    卧室
    1.57
    ห้อง
    1.57
     windowsill
    1.56
    檯
    1.54
     desks
    1.54
     playroom
    1.51
     veranda
    1.50
     downstairs
    1.49
    Act Density 0.399%

    No Known Activations