Add statistical-significance note (z-test, 95% CI, p-values) to results section

Sync README with final paper: 5-config positive-delta highlights, capacity-headroom framing, full reproducibility pointer

Add agentic-SFT warm-up section to model card

Add agentic-SFT warm-up: pythia-410m full FT, Task A (TinyStories), tool-call accuracy 1.000 on 50 probes

Update model card with SLM-RL-Agents naming

Fix num_samples 500 -> 200 to match raw evaluation files

Publish 15 SFT + 15 PPO checkpoints for the SLM-RL-Agent framework

initial commit