Which Demographics do LLMs Default to During Annotation?
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866916763605663744 |
|---|---|
| author | Schäfer, Johannes Combs, Aidan Bagdon, Christopher Li, Jiahui Probol, Nadine Greschner, Lynn Papay, Sean Resendiz, Yarik Menchaca Velutharambath, Aswathy Wührl, Amelie Weber, Sabine Klinger, Roman |
| author_facet | Schäfer, Johannes Combs, Aidan Bagdon, Christopher Li, Jiahui Probol, Nadine Greschner, Lynn Papay, Sean Resendiz, Yarik Menchaca Velutharambath, Aswathy Wührl, Amelie Weber, Sabine Klinger, Roman |
| contents | Demographics and cultural background of annotators influence the labels they assign in text annotation -- for instance, an elderly woman might find it offensive to read a message addressed to a "bro", but a male teenager might find it appropriate. It is therefore important to acknowledge label variations to not under-represent members of a society. Two research directions developed out of this observation in the context of using large language models (LLM) for data annotations, namely (1) studying biases and inherent knowledge of LLMs and (2) injecting diversity in the output by manipulating the prompt with demographic information. We combine these two strands of research and ask the question to which demographics an LLM resorts to when no demographics is given. To answer this question, we evaluate which attributes of human annotators LLMs inherently mimic. Furthermore, we compare non-demographic conditioned prompts and placebo-conditioned prompts (e.g., "you are an annotator who lives in house number 5") to demographics-conditioned prompts ("You are a 45 year old man and an expert on politeness annotation. How do you rate {instance}"). We study these questions for politeness and offensiveness annotations on the POPQUORN data set, a corpus created in a controlled manner to investigate human label variations based on demographics which has not been used for LLM-based analyses so far. We observe notable influences related to gender, race, and age in demographic prompting, which contrasts with previous studies that found no such effects. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_08820 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Which Demographics do LLMs Default to During Annotation? Schäfer, Johannes Combs, Aidan Bagdon, Christopher Li, Jiahui Probol, Nadine Greschner, Lynn Papay, Sean Resendiz, Yarik Menchaca Velutharambath, Aswathy Wührl, Amelie Weber, Sabine Klinger, Roman Computation and Language Demographics and cultural background of annotators influence the labels they assign in text annotation -- for instance, an elderly woman might find it offensive to read a message addressed to a "bro", but a male teenager might find it appropriate. It is therefore important to acknowledge label variations to not under-represent members of a society. Two research directions developed out of this observation in the context of using large language models (LLM) for data annotations, namely (1) studying biases and inherent knowledge of LLMs and (2) injecting diversity in the output by manipulating the prompt with demographic information. We combine these two strands of research and ask the question to which demographics an LLM resorts to when no demographics is given. To answer this question, we evaluate which attributes of human annotators LLMs inherently mimic. Furthermore, we compare non-demographic conditioned prompts and placebo-conditioned prompts (e.g., "you are an annotator who lives in house number 5") to demographics-conditioned prompts ("You are a 45 year old man and an expert on politeness annotation. How do you rate {instance}"). We study these questions for politeness and offensiveness annotations on the POPQUORN data set, a corpus created in a controlled manner to investigate human label variations based on demographics which has not been used for LLM-based analyses so far. We observe notable influences related to gender, race, and age in demographic prompting, which contrasts with previous studies that found no such effects. |
| title | Which Demographics do LLMs Default to During Annotation? |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2410.08820 |