The authors of the article in topic provide a more coherent explanation than I could ever articulate:
>> Conventionally, the formation of chemical bonds is due to a decrease in potential energy (PE), often accompanied by small increases in vibrational zero point energy (ZPE). In principle, this basic mechanism can be completely reversed, wherein chemical bonds may even be formed by an increase in PE if there is a sufficiently compensating decrease in vibrational ZPE, giving rise to what has been coined “vibrational bonding” of molecules stabilized at saddle-point barriers on a potential energy surface (PES), far away from potential minima.
>> Conventionally, the formation of chemical bonds is due to a decrease in potential energy (PE), often accompanied by small increases in vibrational zero point energy (ZPE). In principle, this basic mechanism can be completely reversed, wherein chemical bonds may even be formed by an increase in PE if there is a sufficiently compensating decrease in vibrational ZPE, giving rise to what has been coined “vibrational bonding” of molecules stabilized at saddle-point barriers on a potential energy surface (PES), far away from potential minima.