helloAK96/chaosops-grpo-lora-p3a

Fix model card frontmatter: base_model -> Qwen/Qwen2.5-3B-Instruct

Phase 3A model card: full eval results, training recipe, trade-off notes

Add post-training evaluation.json

Add post-training evaluation_summary.txt

Add post-training comparison_curve.png

Add learning_curve.png

Add training_metrics.json

Upload GRPO LoRA adapter

initial commit