Returned OV 2024.3 references for OV 2024.5 in Weight Compression tes…

…ts (#3065)

### Changes

Revert of #2954
Use [references for the OV
2024.3](https://github.com/ljaljushkin/nncf_pytorch/blob/873fbb80ab07bd41a3b1a2e21f68bc1a7a394810/tests/post_training/data/wc_reference_data.yaml)
when run conformance tests with OV 2024.5.

### Reason for changes

According to the CPU team: 
> 1) In 24.4 time frame, RMS impl has been enabled
(openvinotoolkit/openvino#26147), which
introduce some accuracy change due to vrsqrtss which is done by
approximation.
> 2) In 24.5 time frame, fix has been applied
(openvinotoolkit/openvino#26817) to use vsqrtss
instead. With that we can see the result can be covered to that prior to
openvinotoolkit/openvino#26147
> With that (in 24.5), the accuracy result is recovered to that prior to
RMS impl openvinotoolkit/openvino#26147. Let's
say that 24.4 may has some accuracy issue in some case. 24.5 fix that
issue.

### Related tickets

156605
156776
156237
151260

### Tests

- [x] 47 build of
weekly/job/openvino-nightly/job/post_training_weight_compression

![image](https://github.com/user-attachments/assets/10271e74-4e0b-4506-917a-306481b500e3)
- [x] 48 build for xfail tests

![image](https://github.com/user-attachments/assets/dc5a9424-35fa-4015-965a-ca08fa46e4b7)

tests/post_training/data/wc_reference_data_2024.5.yaml

@@ -1,8 +1,34 @@
 tinyllama_NF4_scale_estimation_stateful_per_channel_backend_OV:
-  metric_value: 0.87132
+  metric_value: 0.88663
   num_int4: 11
   num_int8: 290
+  metrics_xfail_reason: "Issue-148819"
 tinyllama_int4_data_free_backend_TORCH:
   metric_value: 0.73873
   num_int4: 114
   num_int8: 84
+tinyllama_awq_backup_mode_none_backend_OV:
+  metric_value: 0.84783
+  num_int4: 208
+  num_int8: 0
+tinyllama_data_aware_awq_scale_estimation_backend_OV:
+  metric_value: 0.85502
+  num_int4: 94
+  num_int8: 124
+tinyllama_data_aware_awq_scale_estimation_stateful_backend_OV:
+  metric_value: 0.85502
+  num_int4: 94
+  num_int8: 124
+tinyllama_data_aware_awq_stateful_backend_OV:
+  metric_value: 0.85616
+  num_int4: 94
+  num_int8: 124
+tinyllama_data_aware_gptq_scale_estimation_stateful_backend_OV:
+  metric_value: 0.86503
+  num_int4: 94
+  num_int8: 124
+  metrics_xfail_reason: "Issue-148819"
+tinyllama_scale_estimation_per_channel_backend_OV:
+  metric_value: 0.81389
+  num_int4: 188
+  num_int8: 124