Upload test.jsonl (balanced 6,000-query 10-domain dataset)

Upload train.jsonl (balanced 6,000-query 10-domain dataset)

Upload README.md (balanced 6,000-query 10-domain dataset)

Upload README.md (balanced 3,750-query dataset)

Upload test.jsonl (balanced 3,750-query dataset)

Upload train.jsonl (balanced 3,750-query dataset)

Upload README.md (balanced 3,750-query dataset)

Upload test.json (balanced 3,750-query dataset)

Upload train.json (balanced 3,750-query dataset)

Upload README.md (balanced 3,750-query dataset)

Delete stream_routing_benchmark.jsonl with huggingface_hub

Delete stream_routing_benchmark.json with huggingface_hub

Update README.md

Fix dataset viewer: add configs pointing to JSONL split

Add dataset card: description, schema, loading instructions, citation

Add JSONL version for HuggingFace dataset viewer

Add stream_routing_benchmark.json: 6,912 queries across 6 domains x 3 complexity classes, reasoning-depth rubric, κ=0.819

initial commit