Tokenization Matters! Degrading Large Language Models through Challenging Their Tokenization
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_ | 1866915288100896768 |
|---|---|
| author | Wang, Dixuan Li, Yanda Jiang, Junyuan Ding, Zepeng Luo, Ziqin Jiang, Guochao Liang, Jiaqing Yang, Deqing |
| author_facet | Wang, Dixuan Li, Yanda Jiang, Junyuan Ding, Zepeng Luo, Ziqin Jiang, Guochao Liang, Jiaqing Yang, Deqing |
| contents | Large Language Models (LLMs) have shown remarkable capabilities in language understanding and generation. Nonetheless, it was also witnessed that LLMs tend to produce inaccurate responses to specific queries. This deficiency can be traced to the tokenization step LLMs must undergo, which is an inevitable limitation inherent to all LLMs. In fact, incorrect tokenization is the critical point that hinders LLMs in understanding the input precisely, thus leading to unsatisfactory output. This defect is more obvious in Chinese scenarios. To demonstrate this flaw of LLMs, we construct an adversarial dataset, named as $\textbf{ADT (Adversarial Dataset for Tokenizer)}$, which draws upon the vocabularies of various open-source LLMs to challenge LLMs' tokenization. ADT consists of two subsets: the manually constructed ADT-Human and the automatically generated ADT-Auto. Our empirical results reveal that our ADT is highly effective on challenging the tokenization of leading LLMs, including GPT-4o, Llama-3, Deepseek-R1 and so on, thus degrading these LLMs' capabilities. Moreover, our method of automatic data generation has been proven efficient and robust, which can be applied to any open-source LLMs. In this paper, we substantially investigate LLMs' vulnerability in terms of challenging their token segmentation, which will shed light on the subsequent research of improving LLMs' capabilities through optimizing their tokenization process and algorithms. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_17067 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Tokenization Matters! Degrading Large Language Models through Challenging Their Tokenization Wang, Dixuan Li, Yanda Jiang, Junyuan Ding, Zepeng Luo, Ziqin Jiang, Guochao Liang, Jiaqing Yang, Deqing Computation and Language Artificial Intelligence Large Language Models (LLMs) have shown remarkable capabilities in language understanding and generation. Nonetheless, it was also witnessed that LLMs tend to produce inaccurate responses to specific queries. This deficiency can be traced to the tokenization step LLMs must undergo, which is an inevitable limitation inherent to all LLMs. In fact, incorrect tokenization is the critical point that hinders LLMs in understanding the input precisely, thus leading to unsatisfactory output. This defect is more obvious in Chinese scenarios. To demonstrate this flaw of LLMs, we construct an adversarial dataset, named as $\textbf{ADT (Adversarial Dataset for Tokenizer)}$, which draws upon the vocabularies of various open-source LLMs to challenge LLMs' tokenization. ADT consists of two subsets: the manually constructed ADT-Human and the automatically generated ADT-Auto. Our empirical results reveal that our ADT is highly effective on challenging the tokenization of leading LLMs, including GPT-4o, Llama-3, Deepseek-R1 and so on, thus degrading these LLMs' capabilities. Moreover, our method of automatic data generation has been proven efficient and robust, which can be applied to any open-source LLMs. In this paper, we substantially investigate LLMs' vulnerability in terms of challenging their token segmentation, which will shed light on the subsequent research of improving LLMs' capabilities through optimizing their tokenization process and algorithms. |
| title | Tokenization Matters! Degrading Large Language Models through Challenging Their Tokenization |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2405.17067 |