1: \begin{abstract}

2: Polarization issue is generally subject to ideological polarization and affective polarization. In particular, affective polarization usually  accelerates the polarization process and transform social interactions into a zero-sum game. Yet, a wide array of existing

3: literature have not provided valid ways to make distinction between them. Therefore, the mechanism contributing to the rise of affective polarization still remain unclear, as well as its unique emergent dynamics. To address this issue, this study introduces the coupled feedback between opinions and response susceptibility to a attraction-repulsion model which takes account into three parameters: interaction strength, response susceptibility and tolerance to others. The model features phase diagrams of global consensus, affective polarization, and ``harmony with diversity" states. The simulations on time-varying and static social networks show that intermediate parameter ranges yield a global convergence, as one integrated cluster collapsing and converging towards a uncertain moderate position after long-time persistence. Overall, the simulations reveal that the feedback essentially offers a counterforce to establish an inversion between global convergence and ``harmony with diversity". Remarkably, strengthening feedback may facilitate polarization by driving the system priorly self-organize into one integrated cluster which then gradually approaching polarization, especially for low tolerance and strong interactions, and the step-like dynamic behaviors of opinion entropy suggest the occurrence of dynamic equilibrium. The above phenomena have never been unearthed before, which can be regarded as unique dynamics features of affective polarization. For the first time, this study attempts to offer a useful approach to the micro foundations of affective polarization, and the results guide us how to avoid the dilemmas from this polarization.

4: \end{abstract}

5: