Investigating the Influence of Prompt-Specific Shortcuts in AI Generated Text Detection
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_ | 1866929397321170944 |
|---|---|
| author | Park, Choonghyun Kim, Hyuhng Joon Kim, Junyeob Kim, Youna Kim, Taeuk Cho, Hyunsoo Jo, Hwiyeol Lee, Sang-goo Yoo, Kang Min |
| author_facet | Park, Choonghyun Kim, Hyuhng Joon Kim, Junyeob Kim, Youna Kim, Taeuk Cho, Hyunsoo Jo, Hwiyeol Lee, Sang-goo Yoo, Kang Min |
| contents | AI Generated Text (AIGT) detectors are developed with texts from humans and LLMs of common tasks. Despite the diversity of plausible prompt choices, these datasets are generally constructed with a limited number of prompts. The lack of prompt variation can introduce prompt-specific shortcut features that exist in data collected with the chosen prompt, but do not generalize to others. In this paper, we analyze the impact of such shortcuts in AIGT detection. We propose Feedback-based Adversarial Instruction List Optimization (FAILOpt), an attack that searches for instructions deceptive to AIGT detectors exploiting prompt-specific shortcuts. FAILOpt effectively drops the detection performance of the target detector, comparable to other attacks based on adversarial in-context examples. We also utilize our method to enhance the robustness of the detector by mitigating the shortcuts. Based on the findings, we further train the classifier with the dataset augmented by FAILOpt prompt. The augmented classifier exhibits improvements across generation models, tasks, and attacks. Our code will be available at https://github.com/zxcvvxcz/FAILOpt. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_16275 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Investigating the Influence of Prompt-Specific Shortcuts in AI Generated Text Detection Park, Choonghyun Kim, Hyuhng Joon Kim, Junyeob Kim, Youna Kim, Taeuk Cho, Hyunsoo Jo, Hwiyeol Lee, Sang-goo Yoo, Kang Min Computation and Language AI Generated Text (AIGT) detectors are developed with texts from humans and LLMs of common tasks. Despite the diversity of plausible prompt choices, these datasets are generally constructed with a limited number of prompts. The lack of prompt variation can introduce prompt-specific shortcut features that exist in data collected with the chosen prompt, but do not generalize to others. In this paper, we analyze the impact of such shortcuts in AIGT detection. We propose Feedback-based Adversarial Instruction List Optimization (FAILOpt), an attack that searches for instructions deceptive to AIGT detectors exploiting prompt-specific shortcuts. FAILOpt effectively drops the detection performance of the target detector, comparable to other attacks based on adversarial in-context examples. We also utilize our method to enhance the robustness of the detector by mitigating the shortcuts. Based on the findings, we further train the classifier with the dataset augmented by FAILOpt prompt. The augmented classifier exhibits improvements across generation models, tasks, and attacks. Our code will be available at https://github.com/zxcvvxcz/FAILOpt. |
| title | Investigating the Influence of Prompt-Specific Shortcuts in AI Generated Text Detection |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2406.16275 |