PERSONA: A Reproducible Testbed for Pluralistic Alignment
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913443994402816 |
|---|---|
| author | Castricato, Louis Lile, Nathan Rafailov, Rafael Fränken, Jan-Philipp Finn, Chelsea |
| author_facet | Castricato, Louis Lile, Nathan Rafailov, Rafael Fränken, Jan-Philipp Finn, Chelsea |
| contents | The rapid advancement of language models (LMs) necessitates robust alignment with diverse user values. However, current preference optimization approaches often fail to capture the plurality of user opinions, instead reinforcing majority viewpoints and marginalizing minority perspectives. We introduce PERSONA, a reproducible test bed designed to evaluate and improve pluralistic alignment of LMs. We procedurally generate diverse user profiles from US census data, resulting in 1,586 synthetic personas with varied demographic and idiosyncratic attributes. We then generate a large-scale evaluation dataset containing 3,868 prompts and 317,200 feedback pairs obtained from our synthetic personas. Leveraging this dataset, we systematically evaluate LM capabilities in role-playing diverse users, verified through human judges, and the establishment of both a benchmark, PERSONA Bench, for pluralistic alignment approaches as well as an extensive dataset to create new and future benchmarks. The full dataset and benchmarks are available here: https://www.synthlabs.ai/research/persona. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_17387 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | PERSONA: A Reproducible Testbed for Pluralistic Alignment Castricato, Louis Lile, Nathan Rafailov, Rafael Fränken, Jan-Philipp Finn, Chelsea Computation and Language The rapid advancement of language models (LMs) necessitates robust alignment with diverse user values. However, current preference optimization approaches often fail to capture the plurality of user opinions, instead reinforcing majority viewpoints and marginalizing minority perspectives. We introduce PERSONA, a reproducible test bed designed to evaluate and improve pluralistic alignment of LMs. We procedurally generate diverse user profiles from US census data, resulting in 1,586 synthetic personas with varied demographic and idiosyncratic attributes. We then generate a large-scale evaluation dataset containing 3,868 prompts and 317,200 feedback pairs obtained from our synthetic personas. Leveraging this dataset, we systematically evaluate LM capabilities in role-playing diverse users, verified through human judges, and the establishment of both a benchmark, PERSONA Bench, for pluralistic alignment approaches as well as an extensive dataset to create new and future benchmarks. The full dataset and benchmarks are available here: https://www.synthlabs.ai/research/persona. |
| title | PERSONA: A Reproducible Testbed for Pluralistic Alignment |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2407.17387 |