Update README.md (#37)

Author: jaseweston

projects/README.md

@@ -52,6 +52,9 @@ Here we list projects undertaken in the RAM framework that are shared publicly,
 
 ## Evaluation
 
+- **J1** [[paper]](https://arxiv.org/abs/2505.10320).
+  _Learns CoTs for LLM-as-a-Judge via GRPO, outperforms EvalPlanner & Distilled R1 models at 8B and 70B scale._
+
 - **Eval-Planner** [[paper]](https://arxiv.org/abs/2501.18099)).
   _Learning powerful plan+execution CoTs for LLM-as-a-Judge critics, SOTA on RewardBench._