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

    Neuronpedia

    INDEX
    Explanations

    past time references

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     عادة
    0.41
     lifelong
    0.40
    往往
    0.39
    接下來
    0.39
     ensuing
    0.38
     usually
    0.36
     দিচ্ছেন
    0.36
    通常の
    0.35
     typically
    0.34
     newly
    0.33
    POSITIVE LOGITS
    去年
    1.32
     earlier
    1.31
    昨年
    1.27
     yesterday
    1.21
     지난해
    1.19
     years
    1.18
     damals
    1.14
     지난
    1.13
    很久
    1.13
     vuonna
    1.12
    Act Density 0.028%

    No Known Activations