Link the published paper (arXiv:2608.23841); update citation

README.md

@@ -42,7 +42,7 @@ model-index:
 A custom decoder-only transformer with delayed dense FFN + delayed MoE experts,
 designed so its inter-layer dependency graph permits vertical pipelining on CPU.
 Part of the **cflow** project — a CPU-first streaming inference engine written in
-Rust.
+Rust. **Paper: [arXiv:2608.23841](https://arxiv.org/abs/2608.23841).**
 
 > **Hosted weights:** this repository hosts `model.cflow` (17.39 GB) — the
 > **arch2_4_8k_16l** model: 16 layers, hidden 8192, **~31B parameters**
@@ -244,10 +244,16 @@ The cflow project tests 8 claims about CPU inference optimization:
 ## Citation
 
 ```bibtex
-@software{poperszky2026cflow,
-  author = {Poperszky, Tom},
-  title = {cflow: CPU-First Streaming Inference for Pipeline-Native Transformers},
-  year = {2026}
+@misc{poperszky2026cflow,
+  author        = {Poperszky, Tom},
+  title         = {Pipeline-Native Transformers: Co-Designing Model Architecture
+                   and CPU Inference for Bandwidth-Efficient Autoregressive Decode},
+  year          = {2026},
+  eprint        = {2608.23841},
+  archivePrefix = {arXiv},
+  primaryClass  = {cs.LG},
+  url           = {https://arxiv.org/abs/2608.23841}
+}
 }
 ```