Garbage values while using tensor dialect

Hi all, I’m new to MLIR and just started working on it
I experimented few things on tensor dialect in MLIR, when I’m trying to test the IR with tensor.extract_slice operation, I have observed some garbage values getting printed, This is when I wantedly accessed the memory locations which are out of bounds and uninitialized.

Example 1 :

 		%a = arith.constant 1.0 : f32 
                %b = arith.constant 2.0 : f32 
                %c = arith.constant 3.0 : f32 
                %d = arith.constant 4.0 : f32 
                %e = arith.constant 5.0 : f32 
                %f = arith.constant 6.0 : f32 
                %a1 = tensor.from_elements %a,%a,%a,%b,%b,%b,%c,%c,%c,%d,%d,%d,%e,%e,%e,%f,%f,%f : tensor<3x2x3xf32> 
                %b1 = tensor.extract_slice %a1[0,0,0][2,2,3][2,2,1] : tensor<3x2x3xf32> to tensor<2x2x3xf32> 
		
	input = [[[1,1,1],[2,2,2]],[[3,3,3],[4,4,4]],[[5,5,5],[6,6,6]]]

	output = [[[1,    1,    1], [3,    3,    3]], 
		[[5,    5,    5], [5.73972e-42,    0,    0]]] 

Example 2:

%a1 = tensor.from_elements %a,%a,%a,%b,%b,%b,%c,%c,%c,%d,%d,%d,%e,%e,%e,%f,%f,%f : tensor<3x2x3xf32> 
    %b1 = tensor.extract_slice %a1[0,0,0][2,2,3][2,2,2] : tensor<3x2x3xf32> to tensor<2x2x3xf32> 
	
input = [[[1,1,1],[2,2,2]],[[3,3,3],[4,4,4]],[[5,5,5],[6,6,6]]]

output = [[[1,    1,    2], [3,    3,    4]], 
	[[4.6993e+10,    4.35646e+10,    1.68156e-44], [4.35895e+10,    4.35897e+10,    -4.00637e-06]]]

I have tried the same thing with another tensor operation, tensor.extract which also produced garbage value.
Consider Example with same input:

	%a = arith.constant 1.0 : f32 
                %b = arith.constant 2.0 : f32 
                %c = arith.constant 3.0 : f32 
                %d = arith.constant 4.0 : f32 
                %e = arith.constant 5.0 : f32 
                %f = arith.constant 6.0 : f32 
                %a1 = tensor.from_elements %a,%a,%a,%b,%b,%b,%c,%c,%c,%d,%d,%d,%e,%e,%e,%f,%f,%f : tensor<3x2x3xf32> 
		%idx = arith.constant 4 : index
		%val = tensor.extract %a1[idx,idx,idx] : tensor<3x2x3xf32>
	Gives output 0 or any garbage value.

is this is a bug or issue related to memory management?