Add Qwen3.5-35B-A3B OSCAR KV rotations

Qwen3.5-35B-A3B/README.md

@@ -0,0 +1,4 @@
+# Qwen3.5-35B-A3B OSCAR KV rotations
+
+Hybrid linear-attention: 10 full-attention layers (3, 7, ... 39), `head_dim` 256,
+keyed by global layer id. Shared per-layer rotation, `format_version: 1`.

Qwen3.5-35B-A3B/k_rotation_qqt_r_h_pbr.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5fc3ab9da61e7295b5c7fb661a8d776d97861bfcafeefa6bfd121473a99a6148
+size 2638701

Qwen3.5-35B-A3B/v_rotation_sst_r_h_pbr.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:103b0dd6d4b2d6b580609a3f00cef85c974112a3ce3e0df3659234d66eacea20
+size 2638701