TimeBench: A Comprehensive Evaluation of Temporal Reasoning Abilities in Large Language Models

This is the repository containing evaluation datas, instructions and demonstrations with ACL 2024 paper TimeBench: A Comprehensive Evaluation of Temporal Reasoning Abilities in Large Language Models (Chu et al., 2023)

Datasets

Symbolic Temporal Reasoning

• TimeX-NLI (Thukral et al., 2021)
• Date Arithmetic (Tan et al., 2023)

Commonsense Temporal Reasoning

• MCTACO (Zhou et al., 2019)
• TimeDial (Qin et al., 2021)
• DurationQA (Giovanni et al., 2022)
• SituatedGen (Zhang et al, 2023)

Event Temporal Reasoning

• TimeQA (Chen et al., 2021)
• TempReason (Tan et al., 2023)
• MenatQA (Wei et al., 2023)
• TRACIE (Zhou et al., 2021)

Models

• GPT-4 (OpenAI, 2023)
• GPT-3.5 (OpenAI, 2022)
• LLaMA2 (Touvron et al., 2023)
• Baichuan2 (Yang et al., 2023)
• Vicuna-1.5 (Chiang et al., 2023)
• Mistral (Jiang et al., 2023)
• ChatGLM3 (Zeng et al., 2023)
• FLAN-T5 (Chung et al., 2022)

Performance

Citation

If you find our work helpful, you can cite this paper as:

@misc{chu2023timebench,
      title={TimeBench: A Comprehensive Evaluation of Temporal Reasoning Abilities in Large Language Models}, 
      author={Zheng Chu and Jingchang Chen and Qianglong Chen and Weijiang Yu and Haotian Wang and Ming Liu and Bing Qin},
      year={2023},
      eprint={2311.17667},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2311.17667}
}