INDEX
    Explanations

    connect, link

    The neuron detects discourse cues that explicitly signal a logical or rhetorical link between ideas (e.g. “link,” “connection,” “making the connection”).

    New Auto-Interp
    Negative Logits
    257
    -0.07
    	class
    -0.07
    .range
    -0.06
    ۲۷
    -0.06
    _DRAW
    -0.06
    性能
    -0.06
     Omn
    -0.06
     monumental
    -0.06
     GetComponent
    -0.06
    _negative
    -0.06
    POSITIVE LOGITS
     identified
    0.07
     identify
    0.07
     consequently
    0.07
     Moj
    0.07
     ----------
    0.06
    ільш
    0.06
    PI
    0.06
     tracing
    0.06
     Contin
    0.06
     moistur
    0.06
    Act Density 0.033%

    No Known Activations