AEGIS data pipeline
Raw source files belong in data/raw/; normalized records belong in data/processed/.
python data/scripts/preprocess.py data/raw/source.jsonl data/processed/normalized.jsonl --domain math
python data/scripts/generate_debate.py data/processed/normalized.jsonl data/processed/aegis-debate.jsonl
The generator is deterministic and mock-first. Replace its judge/adversary calls with checkpoint inference for research runs, while preserving the trajectory schema consumed by the backend.
Xet Storage Details
- Size:
- 522 Bytes
- Xet hash:
- 70ae06c3d725b7c6ac8cabcbdf17f9900175bd3c96faf636342b9221c0a5641e
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.