INDEX
    Explanations

    The neuron detects descriptive words highlighting accommodations’ comforts and amenities, such as adjectives naming qualities (comfy, clean, fully stocked) and nearby features (walking distance, beach).

    New Auto-Interp
    Negative Logits
    ція
    -0.07
     cin
    -0.06
    increment
    -0.06
     milionů
    -0.06
    (),
    -0.06
    -0.06
     vene
    -0.06
    _epi
    -0.06
     ηλεκ
    -0.06
     formatted
    -0.06
    POSITIVE LOGITS
     jewish
    0.07
     Суд
    0.06
     â
    0.06
     делать
    0.06
    /Q
    0.06
     FOREIGN
    0.06
    是我
    0.06
    щина
    0.06
    ัณฑ
    0.06
     quận
    0.06
    Act Density 0.062%

    No Known Activations