[CPU] Fixed BF16 Matmul inference precision #23312

Job	Run time
Smart_CI	23s
Build	14m 45s
Conformance (OP)	19m 8s
Conformance (API)	2m 14s
NVIDIA plugin	0s
ONNX Runtime Integration	0s
GPU_Stub	0s
OpenVINO tokenizers extension	0s
Debian Packages / Debian Packages	2m 2s
OpenVINO JS API / OpenVINO JS API	1m 5s
Samples / Samples	6m 18s
CPU functional tests / CPU functional tests	17m 52s
C++ unit tests / C++ unit tests	1m 36s
ONNX Models Tests / ONNX Models tests	25m 18s
Python unit tests	0s
TensorFlow Hub Models tests	0s
PyTorch Models tests	0s
ci/gha_overall_status	1s
	1h 30m 42s

Provide feedback