NPUW: LLMInferRequest - not copy kvcache for last generated token #17647

Job	Run time
Smart_CI	24s
Docker	50s
Build / Build	13m 1s
Conformance (OP)	8m 56s
Conformance (API)	2m 12s
NVIDIA plugin	2m 54s
dGPU Tests	0s
iGPU Tests	0s
Debian Packages / Debian Packages	1m 40s
C++ unit tests / C++ unit tests	14m 12s
OpenVINO JS API / OpenVINO JS API	1m 26s
Samples / Samples	4m 40s
JAX Layer Tests / JAX Layer Tests	2m 15s
CPU functional tests / CPU functional tests	23m 4s
ONNX Runtime Integration / ONNX Runtime Integration	6m 37s
ONNX Models Tests / ONNX Models tests	20m 7s
Python unit tests / Python unit tests	4m 13s
Python API tests / Python API tests	2m 21s
Pytorch Layer Tests / PyTorch Layer Tests	15m 46s
OpenVINO tokenizers extension / OpenVINO tokenizers extension	2m 33s
TensorFlow TF Hub Models tests	0s
TensorFlow Hugging Face Models tests	0s
PyTorch Models Nightly Scope1 tests	0s
PyTorch Models Nightly Scope2 tests	0s
TensorFlow Layer Tests / TensorFlow Layer Tests	15m 26s
TensorFlow Models tests / TensorFlow Models tests	13m 18s
PyTorch Models tests / PyTorch Models tests	34m 25s
JAX/Flax Models tests / JAX/Flax Models tests	1m 41s
ci/gha_overall_status	1s
	3h 12m 2s

Provide feedback