README: embed reward curve and belief-accuracy curve plots

README.md

@@ -45,7 +45,11 @@ Plus the gpt-5.4 teacher (the upper-bound reference) hits **0.611 in-dist / 0.62
 
 ![SFT v3 loss](plots/sft_v3_training_loss.png)
 
-The bar comparison (random vs heuristic vs distilled student) is in the **Headline result** table above. Numbers source: `eval_results.json` in the [trained model repo](https://huggingface.co/InosLihka/rhythm-env-meta-trained-sft-v3).
+**Reward curve** — mean per-step env reward over training (real env-replay reward, with ±1 std band). Climbs steadily as the agent learns profile-aware play.
+
+![Reward curve](plots/grpo_iter2_reward_curve.png)
+
+**Baseline vs trained** comparison is in the Headline result table above. Numbers source: `eval_results.json` in the [trained model repo](https://huggingface.co/InosLihka/rhythm-env-meta-trained-sft-v3).
 
 ## Why a Life Simulator?