arXiv cs.AI· Samhita Bollepally, Aurora Sloman-Moll, Takashi Yamauchi, Neeka Motamedrasa·· 3 天前AI 评分22
人类与 LLM 在比喻性语言上的表征相似性对齐研究
Representational Similarity Alignment of Figurative Language in Humans and Large Language Models
AI 导读
研究对比人类与四款指令微调 LLM（GPT-4、Gemma-2-9B、Llama-3.2、Mistral-7B）对 240 句对话在惯用性、讽刺、幽默、情感、成语化和 Gen Z 俚语六项语言特征上的评分，表面判断高度一致，但表征相似性模式存在差异，成语化与 Gen Z 俚语尤为明显。
正文 · 原文
Abstract:Large language models (LLMs) often produce judgments similar to those of humans, but their interpretation of figurative language remains unclear. We compared human participants and four instruction-tuned LLMs (GPT-4, Gemma-2-9B, Llama-3.2, and Mistral-7B) on their ratings of 240 dialogue-based sentences spanning six linguistic traits: conventionality, sarcasm, funny, emotion, idiomacy, and Gen Z slang. While humans and LLMs were strongly aligned at the surface level, differences emerged in representational similarity patterns, particularly in idiomacy and Gen Z slang. GPT-4 showed the closest alignment to humans, though all models differed at the representational level, suggesting that similar outputs may arise from different underlying language representations. We also examined the relationship between model scale and representational similarity using the Gemma model series (Gemma-2-9B-IT, Gemma-3-27B-IT, and Gemma-4-31B-IT), and found that representational similarity patterns were largely unaffected by model size.
| Comments: | 19 pages, 14 figures |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2601.09041 [cs.CL] |
| (or arXiv:2601.09041v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2601.09041 arXiv-issued DOI via DataCite |
Submission history
From: Samhita Bollepally [view email]
[v1]
Wed, 14 Jan 2026 00:13:00 UTC (943 KB)
[v2]
Sun, 11 Oct 2026 01:54:59 UTC (609 KB)
来源：arXiv cs.AI · arxiv.org