vipsehgal/qwen3-8b-jee-sft

Add project report with full training and evaluation details

Remove SDPO references from model card

Add model card with training details and eval results (base vs SFT vs SDPO)

Add SDPO data files (rl_prompts, eval_prompts, judge_config, train)

jee finetune v2: retrained SFT on corrected Phase 2 data (14k examples, mojibake/PhysReason/PhysicsEval fixes)

Upload sdpo_data/train.jsonl with huggingface_hub

Upload sdpo_data/judge_config.json with huggingface_hub

Upload sdpo_data/eval_prompts.jsonl with huggingface_hub

Upload sdpo_data/rl_prompts.jsonl with huggingface_hub

Upload README.md with huggingface_hub

Upload SFT fine-tuned Qwen3-8B for IIT JEE

initial commit