INDEX
    Explanations

    Conversational snippets

    The neuron detects when the user asks about their previous or first message (i.e. meta-questions requesting a recall of earlier conversation).

    New Auto-Interp
    Negative Logits
     herself
    -0.07
    -0.06
    '>↵
    -0.06
    period
    -0.06
     michael
    -0.06
     рассказ
    -0.06
     sözleş
    -0.06
    _current
    -0.06
     Wisdom
    -0.06
     cease
    -0.06
    POSITIVE LOGITS
    itu
    0.07
     तत
    0.07
     ран
    0.06
    0.06
    ีความ
    0.06
     melakukan
    0.06
    orderid
    0.06
    ён
    0.06
     cycl
    0.06
    ový
    0.06
    Act Density 0.032%

    No Known Activations