SonaMath-0.5B: rebrand from JunMath + SFT GSM8K full 13.8%

config: authoritative gsm8k_full_greedy 13.8% (182/1319)

Config: release_stage=sft, rl_applied=false, split research tracks

Add rl_research metadata (GRPO alternative framing)

Add theory_research field to public config

Update config research_focus for SLM reasoning architecture

Release JunMath-0.5B research preview (weights + tokenizer + model card)