ByteDance-Seed/ReSA
ReSA (Reasoned Safety Alignment) Project Page | Paper ReSA (Reasoned Safety Alignment) is an open-source synthetic safety-training dataset with 80K examples designed to enhance LLM robustness against jailbreak attacks through an "Answer-Then-Check" strategy. The dataset teaches models to first generate a summary of their intended answer, then critically evaluate its safety before providing a final response. This approach achieves superior safety performance while maintaining… See the full description on the dataset page: https://huggingface.co/datasets/ByteDance-Seed/ReSA.
| Author / organization | ByteDance-Seed |
|---|---|
| License (as declared) | odc-by |
| Downloads | 166 (as reported by Hugging Face) |
| Likes | 6 |
| Created | 20 October 2025 |
| Last updated at source | 13 March 2026 |
Recent changes
- 16 Oct 2026 Description: ReSA (Reasoned Safety Alignment) Project Page | Paper ReSA (Reasoned Safety Alignment) is an open-source synthetic safety-training dataset with 80K examples designed to enhance LLM robustness against jailbreak attacks through an "Answer-The → ReSA (Reasoned Safety Alignment) Project Page | Paper ReSA (Reasoned Safety Alignment) is an open-source synthetic safety-training dataset with 80K examples designed to enhance LLM robustness against jailbreak attacks through an "Answer-The
Changes detected by Silytics between refreshes of the source.
More from ByteDance-Seed
- ByteDance-Seed/Multi-SWE-benchDataset
- ByteDance-Seed/THEMolDataset
- ByteDance-Seed/WideSearchDataset
- ByteDance-Seed/Multi-SWE-bench_miniDataset
- ByteDance-Seed/EdgeBenchDataset
- ByteDance-Seed/Code-Contests-PlusDataset
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 16 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.