INDEX
    Explanations

    problems, evaluations

    This neuron detects technical phrases expressing research aims or measurement processes, such as “barriers to…,” “detection of…,” “greater control in…,” or “predict the….”

    New Auto-Interp
    Negative Logits
    ็นท
    -0.07
     incorrect
    -0.07
    .leading
    -0.06
     noche
    -0.06
    -0.06
     Pandora
    -0.06
    、あ
    -0.06
     {↵↵
    -0.06
    vironments
    -0.06
    ANTA
    -0.06
    POSITIVE LOGITS
    στη
    0.08
     aslında
    0.07
    trie
    0.06
     descend
    0.06
    .readlines
    0.06
    _BS
    0.06
     vlak
    0.06
     içeri
    0.06
    DWORD
    0.06
    acid
    0.06
    Act Density 0.153%

    No Known Activations