Rot. KE formula considers how the mass of the rod is distributed around the axis of rotation. It accounts for the fact that different parts of the rod are moving at different linear speeds, proportional to their distance from the axis of rotation.
Linear KE formula treats the rod as if all its mass were concentrated at its center of mass, moving at the same speed. It does not account for the rotational motion around the axis.
Like it could work if KE itselft was linear with speed, but since it's not, even if we use speed of center of mass for linear KE formula, those far from the middle cannot properly balance those closer from the middle (because their average KE won't be equal to KE of the middle). So taking full mass and speed of mass center won't give us sum of KE of all particles.
If you want a simply example, think of rod as particles with equal masses. If KE was linear with speed(let's say mv/2), average KE of particles 0.49 L and 0.51 L distances far from center of rotation will be equal to the middle one's. So do 0.48 and 0.52 and so on. (Of course 0.01 difference is just for example) then if we have N particles, and average speed of them all is speed of the middle particle (let's say v), we can say that sum of all their speed is v * N, if from formula we write 1/2 *m * v * N to find sum of all their KE(where m is mass of 1 particle) m * N will be total mass of the rod. As you see using linear KE worked but that won't be true as KE isn't linear with distance over time. In this case it should be linear with distance from axis.
Another one can be when you use average speed to find distance. If speed is increasing linearly, let's say from 0 to v at t time, its average speed will be v/2 at t/2 time so you can just find d = v/2 * t.(it's like that bc just like in example above, average speed of 0.25t and 0.75 t is speed at 0.5t and so on) However same thing won't be true when speed will increase with v=t^2 function as at t/2 time it won't be average speed and we can't find distance from that. What I want to say is just think of t axis as length of rod and v axis as KE from the graphic. In short If we look at all dots in the graph, dividing sum of all their speed by their number won't give us speed at t/2 time.
Sorry if it's complicated but I don't thing I can explain in another way at least for now