Update README.md

README.md

@@ -1,3 +1,33 @@
----
-license: cc-by-sa-4.0
----
+---
+license: cc-by-sa-4.0
+---
+
+# PDDL2PRM
+
+PDDL2PRM is a large-scale dataset for training Process Reward Models (PRMs) with fine-grained step-level supervision derived from PDDL planning problems.
+
+The dataset contains ~1M reasoning steps across 11 domains, where each step is automatically labeled using rule-based rewards reflecting executability, optimality, and progress toward the goal. This enables scalable, precise, and reproducible supervision beyond the mathematical domain.
+
+🔗 **Project page & paper**: https://pisanoraffaele.github.io/prm_meets_planning/
+
+## Key Features
+- ~1M reasoning steps
+- 14,714 planning problems
+- 5-level reward signal (from invalid to optimal)
+- Rule-based annotation
+- Multi-domain (planning, logic, navigation, puzzles)
+
+## Reward Levels
+- 0.0 → Non-executable  
+- 0.25 → Dead-end  
+- 0.5 → Backtracking  
+- 0.75 → Suboptimal  
+- 1.0 → Optimal  
+
+
+## Note
+The data is structurally precise but generated via templates; it is most effective when combined with natural CoT datasets (e.g., PRM800k, Math-Shepherd).
+
+## Citation
+If you use this dataset, please cite:
+Pisano et al., *PRM Meets Planning*