1: \begin{abstract}

2:   We are interested in the high-order approximation of anisotropic\rev{, potential-driven advection-diffusion models} on general polytopal partitions.

3:   We study two hybrid schemes, both built upon the Hybrid High-Order technology.

4:   The first one hinges on exponential fitting and is linear, whereas the second is nonlinear.

5:   The existence of solutions is established for both schemes. Both schemes are also shown to possess a discrete entropy structure, ensuring that the long-time behaviour of discrete solutions mimics the PDE one.

6:   For the nonlinear scheme, the positivity of discrete solutions is a built-in feature.

7:   %% The nonlinear scheme is designed so as to enforce the positivity of discrete solutions.

8:   On the contrary, we display numerical evidence indicating that the linear scheme violates positivity, whatever the order.

9:   Finally, we verify numerically that the nonlinear scheme has optimal order of convergence, expected long-time behaviour, and that raising the polynomial degree results, also in the nonlinear case, in an efficiency gain.

10:   \medskip\\

11:   {\small {\bf Keywords:} High-order methods; Hybrid methods; Polytopal meshes; Structure-preserving schemes; \rev{Potential-driven advection-diffusion}; Long-time behaviour; Entropy methods.}

12:   \smallskip\\

13:   {\small {\bf MSC 2020:} 65M60, 35K51, 35Q84, 35B40.}

14: \end{abstract}

15: