Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2503.10838 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910888011759616 |
|---|---|
| author | Lee, So Young Scheinberg, Russell Shore, Amber Agrawal, Ameeta |
| author_facet | Lee, So Young Scheinberg, Russell Shore, Amber Agrawal, Ameeta |
| contents | This study explores how recent large language models (LLMs) navigate relative clause attachment {ambiguity} and use world knowledge biases for disambiguation in six typologically diverse languages: English, Chinese, Japanese, Korean, Russian, and Spanish. We describe the process of creating a novel dataset -- MultiWho -- for fine-grained evaluation of relative clause attachment preferences in ambiguous and unambiguous contexts. Our experiments with three LLMs indicate that, contrary to humans, LLMs consistently exhibit a preference for local attachment, displaying limited responsiveness to syntactic variations or language-specific attachment patterns. Although LLMs performed well in unambiguous cases, they rigidly prioritized world knowledge biases, lacking the flexibility of human language processing. These findings highlight the need for more diverse, pragmatically nuanced multilingual training to improve LLMs' handling of complex structures and human-like comprehension. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_10838 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Who Relies More on World Knowledge and Bias for Syntactic Ambiguity Resolution: Humans or LLMs? Lee, So Young Scheinberg, Russell Shore, Amber Agrawal, Ameeta Computation and Language This study explores how recent large language models (LLMs) navigate relative clause attachment {ambiguity} and use world knowledge biases for disambiguation in six typologically diverse languages: English, Chinese, Japanese, Korean, Russian, and Spanish. We describe the process of creating a novel dataset -- MultiWho -- for fine-grained evaluation of relative clause attachment preferences in ambiguous and unambiguous contexts. Our experiments with three LLMs indicate that, contrary to humans, LLMs consistently exhibit a preference for local attachment, displaying limited responsiveness to syntactic variations or language-specific attachment patterns. Although LLMs performed well in unambiguous cases, they rigidly prioritized world knowledge biases, lacking the flexibility of human language processing. These findings highlight the need for more diverse, pragmatically nuanced multilingual training to improve LLMs' handling of complex structures and human-like comprehension. |
| title | Who Relies More on World Knowledge and Bias for Syntactic Ambiguity Resolution: Humans or LLMs? |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2503.10838 |