Apply suggestions from code review

Author: liubo-intel

src/plugins/intel_cpu/src/nodes/kernels/scaled_attn/executor_pa.cpp

@@ -1568,7 +1568,7 @@ struct MHA {
                          const PlainTensor& block_indices_begins,
                          const PlainTensor& alibi_slopes,
                          const PlainTensor& score_aggregation_window,
-                         const std::vector<PlainTensor>& sparse_attention_mask = {}) {
+                         const std::vector<PlainTensor>& sparse_attention_mask) {
         auto Hk = v_cache.m_dims[1];
 
         constexpr bool q_is_xf16 = any_of(precision_of<DATA_TYPE>::value, ov::element::bf16, ov::element::f16);