LongBench v2: Towards Deeper Understanding and Reasoning on Realistic Long-context Multitasks
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_ | 1866915089004625920 |
|---|---|
| author | Bai, Yushi Tu, Shangqing Zhang, Jiajie Peng, Hao Wang, Xiaozhi Lv, Xin Cao, Shulin Xu, Jiazheng Hou, Lei Dong, Yuxiao Tang, Jie Li, Juanzi |
| author_facet | Bai, Yushi Tu, Shangqing Zhang, Jiajie Peng, Hao Wang, Xiaozhi Lv, Xin Cao, Shulin Xu, Jiazheng Hou, Lei Dong, Yuxiao Tang, Jie Li, Juanzi |
| contents | This paper introduces LongBench v2, a benchmark designed to assess the ability of LLMs to handle long-context problems requiring deep understanding and reasoning across real-world multitasks. LongBench v2 consists of 503 challenging multiple-choice questions, with contexts ranging from 8k to 2M words, across six major task categories: single-document QA, multi-document QA, long in-context learning, long-dialogue history understanding, code repository understanding, and long structured data understanding. To ensure the breadth and the practicality, we collect data from nearly 100 highly educated individuals with diverse professional backgrounds. We employ both automated and manual review processes to maintain high quality and difficulty, resulting in human experts achieving only 53.7% accuracy under a 15-minute time constraint. Our evaluation reveals that the best-performing model, when directly answers the questions, achieves only 50.1% accuracy. In contrast, the o1-preview model, which includes longer reasoning, achieves 57.7%, surpassing the human baseline by 4%. These results highlight the importance of enhanced reasoning ability and scaling inference-time compute to tackle the long-context challenges in LongBench v2. The project is available at https://longbench2.github.io. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_15204 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | LongBench v2: Towards Deeper Understanding and Reasoning on Realistic Long-context Multitasks Bai, Yushi Tu, Shangqing Zhang, Jiajie Peng, Hao Wang, Xiaozhi Lv, Xin Cao, Shulin Xu, Jiazheng Hou, Lei Dong, Yuxiao Tang, Jie Li, Juanzi Computation and Language Artificial Intelligence This paper introduces LongBench v2, a benchmark designed to assess the ability of LLMs to handle long-context problems requiring deep understanding and reasoning across real-world multitasks. LongBench v2 consists of 503 challenging multiple-choice questions, with contexts ranging from 8k to 2M words, across six major task categories: single-document QA, multi-document QA, long in-context learning, long-dialogue history understanding, code repository understanding, and long structured data understanding. To ensure the breadth and the practicality, we collect data from nearly 100 highly educated individuals with diverse professional backgrounds. We employ both automated and manual review processes to maintain high quality and difficulty, resulting in human experts achieving only 53.7% accuracy under a 15-minute time constraint. Our evaluation reveals that the best-performing model, when directly answers the questions, achieves only 50.1% accuracy. In contrast, the o1-preview model, which includes longer reasoning, achieves 57.7%, surpassing the human baseline by 4%. These results highlight the importance of enhanced reasoning ability and scaling inference-time compute to tackle the long-context challenges in LongBench v2. The project is available at https://longbench2.github.io. |
| title | LongBench v2: Towards Deeper Understanding and Reasoning on Realistic Long-context Multitasks |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2412.15204 |