1: \begin{abstract}

2:   We demonstrate the power of 2D tensor networks for obtaining large deviation functions

3:   of dynamical observables in a classical nonequilibrium setting. 

4: %% A brief overview of the numerical method is provided, 

5: %% with minimal variance from the canonical time-evolving block decimation approach, 

6: %% is followed by a discussion of the method's accuracy and convergence.

7:   %% To emphasize the capabilities of this approach,

8: Using these methods, we analyze the previously unstudied dynamical phase behavior of the 

9: fully 2D asymmetric simple exclusion process with biases in both the $x$ and $y$ directions.

10: We identify a dynamical phase transition, from a jammed to a flowing phase, 

11: and characterize the phases and the transition, with an estimate of the critical point and exponents.

12: \end{abstract}

13: