4 papers · 1 filter
Can LLMs Deceive CLIP? Benchmarking Adversarial Compositionality of Pre-trained Multimodal Representation via Text Updates
Jaewoo Ahn, Heeseung Yun, Dayoon Ko +1
While pre-trained multimodal representations (e.g., CLIP) have shown impressive capabilities, they exhibit significant compositional vulnerabilities leading to counterintuitive jud…
Is a Peeled Apple Still Red? Evaluating LLMs' Ability for Conceptual Combination with Property Type
Seokwon Song, Taehyun Lee, Jaewoo Ahn +2
Conceptual combination is a cognitive process that merges basic concepts, enabling the creation of complex expressions. During this process, the properties of combination (e.g., th…
Who Wrote this Code? Watermarking for Code Generation
Taehyun Lee, Seokhee Hong, Jaewoo Ahn +5
Since the remarkable generation performance of large language models raised ethical and legal concerns, approaches to detect machine-generated text by embedding watermarks are bein…
TimeChara: Evaluating Point-in-Time Character Hallucination of Role-Playing Large Language Models
Jaewoo Ahn, Taehyun Lee, Junyoung Lim +4
While Large Language Models (LLMs) can serve as agents to simulate human behaviors (i.e., role-playing agents), we emphasize the importance of point-in-time role-playing. This situ…