Update README.md

README.md

@@ -76,19 +76,17 @@ tok = AutoTokenizer.from_pretrained(p, trust_remote_code=True)
 | `model_680m_6.4B` | 680m | 6.4B | 10 | 100–1000 |
 | `model_680m_16B` | 680m | 16B | 4 | 1200–1900 |
 
-## Not included
+## Citation
 
-These models exist in the study but had no usable weights available at upload
-time:
+If you use these models, please cite it:
 
-- `model_200m_0.53B`
-- `model_200m_11B`
-- `model_200m_21B`
-- `model_200m_40B`
-- `model_200m_53B`
-- `model_680m_0.64B`
-- `model_680m_1.2B`
-- `model_680m_12B`
-- `model_680m_3.2B`
-- `model_680m_32B`
-- 2 32m models absent from the pretraining-token tables
+```bibtex
+@article{pre2post-chess,
+  title   = {Understanding Reasoning from Pretraining to Post-Training},
+  author  = {Shen, Jingyan and Li, Ang and Rahman, Salman and Sun, Yifan and
+             Goldblum, Micah and Telgarsky, Matus and Izmailov, Pavel},
+  journal = {arXiv preprint arXiv:2607.16097},
+  year    = {2026},
+  url     = {https://arxiv.org/pdf/2607.16097}
+}
+```