The Essence of Contextual Understanding in Theory of Mind: A Study on Question Answering with Story Characters
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912318019862528 |
|---|---|
| author | Zhou, Chulun Wang, Qiujing Yu, Mo Yue, Xiaoqian Lu, Rui Li, Jiangnan Zhou, Yifan Zhang, Shunchi Zhou, Jie Lam, Wai |
| author_facet | Zhou, Chulun Wang, Qiujing Yu, Mo Yue, Xiaoqian Lu, Rui Li, Jiangnan Zhou, Yifan Zhang, Shunchi Zhou, Jie Lam, Wai |
| contents | Theory-of-Mind (ToM) is a fundamental psychological capability that allows humans to understand and interpret the mental states of others. Humans infer others' thoughts by integrating causal cues and indirect clues from broad contextual information, often derived from past interactions. In other words, human ToM heavily relies on the understanding about the backgrounds and life stories of others. Unfortunately, this aspect is largely overlooked in existing benchmarks for evaluating machines' ToM capabilities, due to their usage of short narratives without global context, especially personal background of characters. In this paper, we verify the importance of comprehensive contextual understanding about personal backgrounds in ToM and assess the performance of LLMs in such complex scenarios. To achieve this, we introduce CharToM benchmark, comprising 1,035 ToM questions based on characters from classic novels. Our human study reveals a significant disparity in performance: the same group of educated participants performs dramatically better when they have read the novels compared to when they have not. In parallel, our experiments on state-of-the-art LLMs, including the very recent o1 and DeepSeek-R1 models, show that LLMs still perform notably worse than humans, despite that they have seen these stories during pre-training. This highlights the limitations of current LLMs in capturing the nuanced contextual information required for ToM reasoning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_01705 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | The Essence of Contextual Understanding in Theory of Mind: A Study on Question Answering with Story Characters Zhou, Chulun Wang, Qiujing Yu, Mo Yue, Xiaoqian Lu, Rui Li, Jiangnan Zhou, Yifan Zhang, Shunchi Zhou, Jie Lam, Wai Computation and Language Artificial Intelligence Theory-of-Mind (ToM) is a fundamental psychological capability that allows humans to understand and interpret the mental states of others. Humans infer others' thoughts by integrating causal cues and indirect clues from broad contextual information, often derived from past interactions. In other words, human ToM heavily relies on the understanding about the backgrounds and life stories of others. Unfortunately, this aspect is largely overlooked in existing benchmarks for evaluating machines' ToM capabilities, due to their usage of short narratives without global context, especially personal background of characters. In this paper, we verify the importance of comprehensive contextual understanding about personal backgrounds in ToM and assess the performance of LLMs in such complex scenarios. To achieve this, we introduce CharToM benchmark, comprising 1,035 ToM questions based on characters from classic novels. Our human study reveals a significant disparity in performance: the same group of educated participants performs dramatically better when they have read the novels compared to when they have not. In parallel, our experiments on state-of-the-art LLMs, including the very recent o1 and DeepSeek-R1 models, show that LLMs still perform notably worse than humans, despite that they have seen these stories during pre-training. This highlights the limitations of current LLMs in capturing the nuanced contextual information required for ToM reasoning. |
| title | The Essence of Contextual Understanding in Theory of Mind: A Study on Question Answering with Story Characters |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2501.01705 |