Summary of Analyzing Cultural Representations Of Emotions in Llms Through Mixed Emotion Survey, by Shiran Dudy et al.
Analyzing Cultural Representations of Emotions in LLMs through Mixed Emotion Survey
by Shiran Dudy, Ibrahim Said Ahmad, Ryoko Kitajima, Agata Lapedriza
First submitted to arxiv on: 4 Aug 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This study investigates the cultural representations of emotions in Large Language Models (LLMs), specifically in mixed-emotion situations. The authors administered a survey used by Miyamoto et al. (2010) to five LLMs and analyzed their outputs. They then experimented with contextual variables to explore variations in responses considering language and speaker origin. The study finds that LLMs have limited alignment with evidence from the literature, and written language has a greater effect on responses than information about participants’ origin. Notably, LLMs’ responses were found to be more similar for East Asian languages compared to Western European languages. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research explores how Large Language Models (LLMs) understand emotions in different cultures. The study uses surveys and experiments to compare how well LLMs represent emotional expressions from various countries, such as Japan and the US. The results show that LLMs are not very good at understanding cultural differences in emotions and tend to rely more on written language than information about where people come from. Interestingly, LLMs were found to be better at representing emotions for languages spoken in East Asia compared to Western Europe. |
Keywords
» Artificial intelligence » Alignment
