Abstract
Recent applications of LLMs in Machine Reading Comprehension (MRC) systems have shown impressive results, but the use of shortcuts, mechanisms triggered by features spuriously correlated to the true label, has emerged as a potential threat to their reliability. We analyze the problem from two angles: LLMs as editors, guided to edit text to mislead LLMs; and LLMs as readers, who answer questions based on the edited text. We introduce a framework that guides an editor to add potential shortcuts-triggers to samples. Using GPT4 as the editor, we find it can successfully edit trigger shortcut in samples that fool LLMs. Analysing LLMs as readers, we observe that even capable LLMs can be deceived using shortcut knowledge. Strikingly, we discover that GPT4 can be deceived by its own edits (15% drop in F1). Our findings highlight inherent vulnerabilities of LLMs to shortcut manipulations. We publish ShortcutQA, a curated dataset generated by our framework for future research.- Anthology ID:
- 2023.findings-emnlp.569
- Volume:
- Findings of the Association for Computational Linguistics: EMNLP 2023
- Month:
- December
- Year:
- 2023
- Address:
- Singapore
- Editors:
- Houda Bouamor, Juan Pino, Kalika Bali
- Venue:
- Findings
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 8495–8505
- Language:
- URL:
- https://aclanthology.org/2023.findings-emnlp.569
- DOI:
- 10.18653/v1/2023.findings-emnlp.569
- Cite (ACL):
- Mosh Levy, Shauli Ravfogel, and Yoav Goldberg. 2023. Guiding LLM to Fool Itself: Automatically Manipulating Machine Reading Comprehension Shortcut Triggers. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 8495–8505, Singapore. Association for Computational Linguistics.
- Cite (Informal):
- Guiding LLM to Fool Itself: Automatically Manipulating Machine Reading Comprehension Shortcut Triggers (Levy et al., Findings 2023)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-5/2023.findings-emnlp.569.pdf