ContinuousBench: Can Differentially Private Synthetic Text Improve Capabilities?
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914622166007808 |
|---|---|
| author | Liu, Peihan Rosenblatt, Lucas Kong, Weiwei Ponomareva, Natalia Kamath, Gautam Cummings, Rachel Geambasu, Roxana Gan, Yu Tsai, Lillian Bie, Alex |
| author_facet | Liu, Peihan Rosenblatt, Lucas Kong, Weiwei Ponomareva, Natalia Kamath, Gautam Cummings, Rachel Geambasu, Roxana Gan, Yu Tsai, Lillian Bie, Alex |
| contents | Differentially private (DP) text synthesis promises to unlock sensitive corpora for model training, but it remains unclear whether DP synthetic data transmits genuinely new knowledge and capabilities present only in those corpora. This is because existing evaluations rely on tasks that are nearly solvable without training, so strong benchmark performance does not establish that DP synthesis can substitute original data access. Thus, we introduce ContinuousBench, a continuously and automatically-regenerated benchmark that measures capability gain from DP synthetic text. Each quarter, a new release pairs a never-before-seen training corpus with a derived QA set, constructed to be: (1) unsolvable sans-corpus; and (2) learnable under DP, as the tested knowledge is supported by hundreds of independent records. Researchers produce DP synthetic data from the training corpus and run our standardized training and evaluation harness on their synthetic data to measure gains. We instantiate two tracks: Geminon, a procedurally-generated dataset about fictional creatures; and News, a stream of newly crawled public news articles. Although standard benchmarks are nearly saturated, on ContinuousBench we find that non-private synthesis transfers substantial knowledge from the original corpus, while state-of-the-art DP synthesis methods generally fail to do so, even at $\varepsilon=100$. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2606_01849 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ContinuousBench: Can Differentially Private Synthetic Text Improve Capabilities? Liu, Peihan Rosenblatt, Lucas Kong, Weiwei Ponomareva, Natalia Kamath, Gautam Cummings, Rachel Geambasu, Roxana Gan, Yu Tsai, Lillian Bie, Alex Machine Learning Computation and Language Cryptography and Security Differentially private (DP) text synthesis promises to unlock sensitive corpora for model training, but it remains unclear whether DP synthetic data transmits genuinely new knowledge and capabilities present only in those corpora. This is because existing evaluations rely on tasks that are nearly solvable without training, so strong benchmark performance does not establish that DP synthesis can substitute original data access. Thus, we introduce ContinuousBench, a continuously and automatically-regenerated benchmark that measures capability gain from DP synthetic text. Each quarter, a new release pairs a never-before-seen training corpus with a derived QA set, constructed to be: (1) unsolvable sans-corpus; and (2) learnable under DP, as the tested knowledge is supported by hundreds of independent records. Researchers produce DP synthetic data from the training corpus and run our standardized training and evaluation harness on their synthetic data to measure gains. We instantiate two tracks: Geminon, a procedurally-generated dataset about fictional creatures; and News, a stream of newly crawled public news articles. Although standard benchmarks are nearly saturated, on ContinuousBench we find that non-private synthesis transfers substantial knowledge from the original corpus, while state-of-the-art DP synthesis methods generally fail to do so, even at $\varepsilon=100$. |
| title | ContinuousBench: Can Differentially Private Synthetic Text Improve Capabilities? |
| topic | Machine Learning Computation and Language Cryptography and Security |
| url | https://arxiv.org/abs/2606.01849 |