diff --git a/.github/workflows/e2e_core.yml b/.github/workflows/e2e_core.yml
index 148232bd0a..32b8d58e7a 100644
--- a/.github/workflows/e2e_core.yml
+++ b/.github/workflows/e2e_core.yml
@@ -169,10 +169,6 @@ jobs:
         -DCMAKE_CXX_COMPILER="$(which clang++)"
         -DLLVM_LIT="${{github.workspace}}/sycl-repo/llvm/utils/lit/lit.py"
 
-    - name: Set LIT_XFAIL_NOT
-      if: inputs.xfail_not != ''
-      run: echo "LIT_XFAIL_NOT=${{inputs.xfail_not}}" >> $GITHUB_ENV
-
     - name: Set LIT_XFAIL
       if: inputs.xfail != ''
       run: echo "LIT_XFAIL=${{inputs.xfail}}" >> $GITHUB_ENV
@@ -181,6 +177,10 @@ jobs:
       if: inputs.filter_out != ''
       run: echo "LIT_FILTER_OUT=${{inputs.filter_out}}" >> $GITHUB_ENV
 
+    - name: Set LIT_XFAIL_NOT
+      if: inputs.xfail_not != ''
+      run: echo "LIT_XFAIL_NOT=${{inputs.xfail_not}}" >> $GITHUB_ENV
+
     # TODO: remove once intel/llvm lit tests can properly recognize the GPU
     - name: Configure hardware platform feature for L0
       if: matrix.adapter.name == 'L0'
diff --git a/.github/workflows/e2e_cuda.yml b/.github/workflows/e2e_cuda.yml
index 6b4b0ca16c..c2f1d969b8 100644
--- a/.github/workflows/e2e_cuda.yml
+++ b/.github/workflows/e2e_cuda.yml
@@ -21,3 +21,4 @@ jobs:
       config: "--cuda"
       unit: "gpu"
       extra_lit_flags: "-sv --max-time=3600"
+      xfail: "Regression/device_num.cpp"
diff --git a/.github/workflows/e2e_level_zero.yml b/.github/workflows/e2e_level_zero.yml
index ecab9a85ba..3b2511e0f9 100644
--- a/.github/workflows/e2e_level_zero.yml
+++ b/.github/workflows/e2e_level_zero.yml
@@ -21,9 +21,11 @@ jobs:
       config: ""
       unit: "gpu"
       # Failing tests
-      xfail: "DeviceCodeSplit/grf.cpp;ESIMD/grf.cpp;ESIMD/mask_expand_load.cpp;KernelAndProgram/target_register_alloc_mode.cpp;Matrix/SG32/get_coord_int8_matB.cpp;Matrix/get_coord_int8_matB.cpp;Matrix/joint_matrix_prefetch.cpp;Matrix/joint_matrix_rowmajorA_rowmajorB.cpp;  SYCL :: ESIMD/mask_expand_load.cpp;Matrix/SPVCooperativeMatrix/joint_matrix_bf16_fill_k_cache_OOB.cpp;Matrix/SPVCooperativeMatrix/joint_matrix_out_bounds.cpp;Matrix/SPVCooperativeMatrix/joint_matrix_prefetch.cpp;Matrix/SPVCooperativeMatrix/joint_matrix_rowmajorA_rowmajorB.cpp;Matrix/element_wise_all_ops_1d.cpp;Matrix/element_wise_all_ops_1d_cont.cpp;Matrix/element_wise_all_ops_scalar.cpp;Matrix/joint_matrix_bf16_fill_k_cache_OOB.cpp;Matrix/joint_matrix_bf16_fill_k_cache_prefetch.cpp;Matrix/joint_matrix_out_bounds.cpp;Matrix/joint_matrix_unaligned_k.cpp"
+      xfail: "DeviceCodeSplit/grf.cpp;ESIMD/mask_expand_load.cpp;KernelAndProgram/target_register_alloc_mode.cpp;Matrix/SG32/get_coord_int8_matB.cpp;Matrix/get_coord_int8_matB.cpp;Matrix/joint_matrix_prefetch.cpp;Matrix/joint_matrix_rowmajorA_rowmajorB.cpp;ESIMD/mask_expand_load.cpp;Matrix/SPVCooperativeMatrix/joint_matrix_bf16_fill_k_cache_OOB.cpp;Matrix/SPVCooperativeMatrix/joint_matrix_out_bounds.cpp;Matrix/SPVCooperativeMatrix/joint_matrix_prefetch.cpp;Matrix/SPVCooperativeMatrix/joint_matrix_rowmajorA_rowmajorB.cpp;Matrix/element_wise_all_ops_1d.cpp;Matrix/element_wise_all_ops_1d_cont.cpp;Matrix/element_wise_all_ops_scalar.cpp;Matrix/joint_matrix_bf16_fill_k_cache_OOB.cpp;Matrix/joint_matrix_bf16_fill_k_cache_prefetch.cpp;Matrix/joint_matrix_out_bounds.cpp;Matrix/joint_matrix_unaligned_k.cpp;Matrix/SPVCooperativeMatrix/SG32/get_coord_int8_matB.cpp;Matrix/SPVCooperativeMatrix/element_wise_all_ops_1d.cpp;Matrix/SPVCooperativeMatrix/element_wise_all_ops_1d_cont.cpp;Matrix/SPVCooperativeMatrix/element_wise_all_ops_scalar.cpp;Matrix/SPVCooperativeMatrix/element_wise_ops.cpp;Matrix/SPVCooperativeMatrix/get_coord_int8_matB.cpp;Matrix/SPVCooperativeMatrix/joint_matrix_bf16_fill_k_cache_SLM.cpp;Matrix/joint_matrix_bf16_fill_k_cache_SLM.cpp"
+      # Unexpectedly Passed Tests
+      xfail_not: ""
       # Flaky tests
-      filter_out: "ESIMD/named_barriers/loop_extended.cpp;ESIMD/local_accessor_copy_to_from.cpp;"
+      filter_out: "Basic/accessor/accessor.cpp|DeviceArchitecture/device_architecture_comparison_on_device_aot.cpp|Graph/Explicit/interop-level-zero-launch-kernel.cpp|Graph/RecordReplay/interop-level-zero-launch-kernel.cpp|syclcompat/launch/launch_policy_lmem.cpp"
       # These runners by default spawn upwards of 260 workers.
       # We also add a time out just in case some test hangs
       extra_lit_flags: "--param gpu-intel-pvc=True --param gpu-intel-pvc-1T=True -sv -j 100 --max-time=3600"
diff --git a/.github/workflows/e2e_opencl.yml b/.github/workflows/e2e_opencl.yml
index fa7984bb3f..e4714b2434 100644
--- a/.github/workflows/e2e_opencl.yml
+++ b/.github/workflows/e2e_opencl.yml
@@ -20,5 +20,5 @@ jobs:
       prefix: ""
       config: ""
       unit: "cpu"
-      xfail: "AOT/double.cpp;AOT/half.cpp;AOT/reqd-sg-size.cpp;Basic/built-ins/marray_geometric.cpp;KernelCompiler/kernel_compiler_spirv.cpp;KernelCompiler/opencl_queries.cpp"
+      xfail: "AOT/double.cpp;AOT/half.cpp;AOT/reqd-sg-size.cpp;Basic/built-ins/marray_geometric.cpp;KernelCompiler/kernel_compiler_spirv.cpp;KernelCompiler/opencl_queries.cpp;NonUniformGroups/ballot_group.cpp;NonUniformGroups/ballot_group_algorithms.cpp;NonUniformGroups/fixed_size_group_algorithms.cpp;NonUniformGroups/opportunistic_group.cpp;NonUniformGroups/opportunistic_group_algorithms.cpp;NonUniformGroups/tangle_group.cpp;NonUniformGroups/tangle_group_algorithms.cpp"
       extra_lit_flags: "-sv --max-time=3600"