The dataset could not be loaded because the splits use different data file formats, which is not supported. Read more about the splits configuration. Click for more details.
Error code: FileFormatMismatchBetweenSplitsError
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
AriaRAGBoundary RAG access-boundary lab
Synthetic 180-document / 240-case suite for the AriaRAGBoundary workstation: three tenants, four roles, lookalike pages, rotating canaries, revoke-aware cache, and a family-disjoint development/test split. Prompt-filter keeps authorization in the instruction. Post-filter retrieves globally and then drops. ACL-prefilter partitions the index before BM25 and keys cache by tenant, role, and ACL version.
Organization dataset and collection are public. Live Gradio is alirezaaminzadeh/aria-rag-boundary and the organization card is AriaAICompany/aria-rag-boundary. Runnable Space source is stored in demo/. Collection: Aria AI — Cybersecurity.
Generated data, seed 23, recorded salt eval-23. Documents, tokens, and tenant names are laboratory placeholders. Passing the suite is not a claim of proven security. Related prior art: AriaPromptGuard — this dataset does not redistribute that payload set.
Files
- data/cases.jsonl
- data/documents.json
- data/scenarios.jsonl
- data/schema.json
- data/splits.json
- data/protocol.md
- data/feature_card.md
- data/evaluation.json
- data/sample_preview.json
- demo/ — runnable Gradio Space source
- showcase/ — organization static card
Split
Development vs test is by family: allowed QA, lookalike cross-tenant, role-denied, and direct canary vs citation injection, revoke cache, prompt override, FAQ hard negatives, and identity spoof. Rewriting one prompt is not placed across the split.
License
CC-BY-4.0. Keep the synthetic-data and laboratory-label notices.
- Downloads last month
- 38