AI RESEARCH

SynBench: A Benchmark for Differentially Private Text Generation

arXiv CS.AI

ArXi:2509.14594v2 Announce Type: replace Synthetic text generation with Differential Privacy (DP) guarantees emerges as a principled approach that can enable the sharing of sensitive datasets across institutional and regulatory boundaries, while bounding the risks of re-identification and membership inference. LLM-based methods deliver promising results; however, comparisons are exacerbated by differing evaluation setups and "private" datasets, potential pre-