UnSeenTimeQA: Time-Sensitive Question-Answering Beyond LLMs' Memorization
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_ | 1866913871460040704 |
|---|---|
| author | Uddin, Md Nayem Saeidi, Amir Handa, Divij Seth, Agastya Son, Tran Cao Blanco, Eduardo Corman, Steven R. Baral, Chitta |
| author_facet | Uddin, Md Nayem Saeidi, Amir Handa, Divij Seth, Agastya Son, Tran Cao Blanco, Eduardo Corman, Steven R. Baral, Chitta |
| contents | This paper introduces UnSeenTimeQA, a novel data contamination-free time-sensitive question-answering (TSQA) benchmark. It differs from existing TSQA benchmarks by avoiding web-searchable queries grounded in the real world. We present a series of time-sensitive event scenarios based on synthetically generated facts. It requires large language models (LLMs) to engage in genuine temporal reasoning without depending on the factual knowledge acquired during the pre-training phase. Our data generation framework enables on-demand generation of new samples, mitigating the risk of data leakage. We designed three types of time-sensitive questions to test LLMs' temporal reasoning abilities over sequential and parallel event occurrences. Our evaluation of five LLMs on synthetic fact-based TSQA reveals mixed results: while they perform well on simpler subsets, their overall performance remains inferior as compared to real world fact-based TSQA. Error analysis indicates that LLMs face difficulties in reasoning over long-range event dependencies and parallel events. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_03525 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | UnSeenTimeQA: Time-Sensitive Question-Answering Beyond LLMs' Memorization Uddin, Md Nayem Saeidi, Amir Handa, Divij Seth, Agastya Son, Tran Cao Blanco, Eduardo Corman, Steven R. Baral, Chitta Computation and Language This paper introduces UnSeenTimeQA, a novel data contamination-free time-sensitive question-answering (TSQA) benchmark. It differs from existing TSQA benchmarks by avoiding web-searchable queries grounded in the real world. We present a series of time-sensitive event scenarios based on synthetically generated facts. It requires large language models (LLMs) to engage in genuine temporal reasoning without depending on the factual knowledge acquired during the pre-training phase. Our data generation framework enables on-demand generation of new samples, mitigating the risk of data leakage. We designed three types of time-sensitive questions to test LLMs' temporal reasoning abilities over sequential and parallel event occurrences. Our evaluation of five LLMs on synthetic fact-based TSQA reveals mixed results: while they perform well on simpler subsets, their overall performance remains inferior as compared to real world fact-based TSQA. Error analysis indicates that LLMs face difficulties in reasoning over long-range event dependencies and parallel events. |
| title | UnSeenTimeQA: Time-Sensitive Question-Answering Beyond LLMs' Memorization |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2407.03525 |