Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2408.01091 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866929448755920896 |
|---|---|
| author | Gao, Jin Gan, Lei Li, Yuankai Ye, Yixin Wang, Dequan |
| author_facet | Gao, Jin Gan, Lei Li, Yuankai Ye, Yixin Wang, Dequan |
| contents | Large multimodal models (LMMs) excel in adhering to human instructions. However, self-contradictory instructions may arise due to the increasing trend of multimodal interaction and context length, which is challenging for language beginners and vulnerable populations. We introduce the Self-Contradictory Instructions benchmark to evaluate the capability of LMMs in recognizing conflicting commands. It comprises 20,000 conflicts, evenly distributed between language and vision paradigms. It is constructed by a novel automatic dataset creation framework, which expedites the process and enables us to encompass a wide range of instruction forms. Our comprehensive evaluation reveals current LMMs consistently struggle to identify multimodal instruction discordance due to a lack of self-awareness. Hence, we propose the Cognitive Awakening Prompting to inject cognition from external, largely enhancing dissonance detection. The dataset and code are here: https://selfcontradiction.github.io/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2408_01091 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Dissecting Dissonance: Benchmarking Large Multimodal Models Against Self-Contradictory Instructions Gao, Jin Gan, Lei Li, Yuankai Ye, Yixin Wang, Dequan Artificial Intelligence Large multimodal models (LMMs) excel in adhering to human instructions. However, self-contradictory instructions may arise due to the increasing trend of multimodal interaction and context length, which is challenging for language beginners and vulnerable populations. We introduce the Self-Contradictory Instructions benchmark to evaluate the capability of LMMs in recognizing conflicting commands. It comprises 20,000 conflicts, evenly distributed between language and vision paradigms. It is constructed by a novel automatic dataset creation framework, which expedites the process and enables us to encompass a wide range of instruction forms. Our comprehensive evaluation reveals current LMMs consistently struggle to identify multimodal instruction discordance due to a lack of self-awareness. Hence, we propose the Cognitive Awakening Prompting to inject cognition from external, largely enhancing dissonance detection. The dataset and code are here: https://selfcontradiction.github.io/. |
| title | Dissecting Dissonance: Benchmarking Large Multimodal Models Against Self-Contradictory Instructions |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2408.01091 |