[Bug] When compiling using TVM's Relax, the output remains the same regardless of the optimization level set.

When compiling an ONNX model using TVM's Relax, the generated .so files are identical when setting opt_level=0, opt_level=1, opt_level=3, and opt_level=4.
The compilation code I used is as follows:

`onnx_model = onnx.load(onnx_path)
onnx_name = os.path.splitext(os.path.basename(onnx_path))[0]
print(onnx_name)

from tvm.relax.frontend.onnx import from_onnx
mod = from_onnx(onnx_model)

with tvm.transform.PassContext(opt_level=0):
ex = relax.build(mod, target="llvm")

libpath = os.path.join(base_path, f"{onnx_name}_relax_O0.so")`

Start with the provided ONNX-to-Relax compilation snippet and reproduce the builds at opt_level 0, 1, 3, and 4. Compare the generated .so files and trace how PassContext is applied during relax.build; the issue is complete when the cause of identical outputs is identified and the expected optimization-level behavior is verified.

Written by the indexing model from the issue text.