Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2501.03200 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909449018408960 |
|---|---|
| author | Jacovi, Alon Wang, Andrew Alberti, Chris Tao, Connie Lipovetz, Jon Olszewska, Kate Haas, Lukas Liu, Michelle Keating, Nate Bloniarz, Adam Saroufim, Carl Fry, Corey Marcus, Dror Kukliansky, Doron Tomar, Gaurav Singh Swirhun, James Xing, Jinwei Wang, Lily Gurumurthy, Madhu Aaron, Michael Ambar, Moran Fellinger, Rachana Wang, Rui Zhang, Zizhao Goldshtein, Sasha Das, Dipanjan |
| author_facet | Jacovi, Alon Wang, Andrew Alberti, Chris Tao, Connie Lipovetz, Jon Olszewska, Kate Haas, Lukas Liu, Michelle Keating, Nate Bloniarz, Adam Saroufim, Carl Fry, Corey Marcus, Dror Kukliansky, Doron Tomar, Gaurav Singh Swirhun, James Xing, Jinwei Wang, Lily Gurumurthy, Madhu Aaron, Michael Ambar, Moran Fellinger, Rachana Wang, Rui Zhang, Zizhao Goldshtein, Sasha Das, Dipanjan |
| contents | We introduce FACTS Grounding, an online leaderboard and associated benchmark that evaluates language models' ability to generate text that is factually accurate with respect to given context in the user prompt. In our benchmark, each prompt includes a user request and a full document, with a maximum length of 32k tokens, requiring long-form responses. The long-form responses are required to be fully grounded in the provided context document while fulfilling the user request. Models are evaluated using automated judge models in two phases: (1) responses are disqualified if they do not fulfill the user request; (2) they are judged as accurate if the response is fully grounded in the provided document. The automated judge models were comprehensively evaluated against a held-out test-set to pick the best prompt template, and the final factuality score is an aggregate of multiple judge models to mitigate evaluation bias. The FACTS Grounding leaderboard will be actively maintained over time, and contains both public and private splits to allow for external participation while guarding the integrity of the leaderboard. It can be found at https://www.kaggle.com/facts-leaderboard. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_03200 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | The FACTS Grounding Leaderboard: Benchmarking LLMs' Ability to Ground Responses to Long-Form Input Jacovi, Alon Wang, Andrew Alberti, Chris Tao, Connie Lipovetz, Jon Olszewska, Kate Haas, Lukas Liu, Michelle Keating, Nate Bloniarz, Adam Saroufim, Carl Fry, Corey Marcus, Dror Kukliansky, Doron Tomar, Gaurav Singh Swirhun, James Xing, Jinwei Wang, Lily Gurumurthy, Madhu Aaron, Michael Ambar, Moran Fellinger, Rachana Wang, Rui Zhang, Zizhao Goldshtein, Sasha Das, Dipanjan Computation and Language We introduce FACTS Grounding, an online leaderboard and associated benchmark that evaluates language models' ability to generate text that is factually accurate with respect to given context in the user prompt. In our benchmark, each prompt includes a user request and a full document, with a maximum length of 32k tokens, requiring long-form responses. The long-form responses are required to be fully grounded in the provided context document while fulfilling the user request. Models are evaluated using automated judge models in two phases: (1) responses are disqualified if they do not fulfill the user request; (2) they are judged as accurate if the response is fully grounded in the provided document. The automated judge models were comprehensively evaluated against a held-out test-set to pick the best prompt template, and the final factuality score is an aggregate of multiple judge models to mitigate evaluation bias. The FACTS Grounding leaderboard will be actively maintained over time, and contains both public and private splits to allow for external participation while guarding the integrity of the leaderboard. It can be found at https://www.kaggle.com/facts-leaderboard. |
| title | The FACTS Grounding Leaderboard: Benchmarking LLMs' Ability to Ground Responses to Long-Form Input |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2501.03200 |