Can LLMs Beat Humans on Discerning Human-written and LLM-generated Science News?The paper introduces SANews, a manually annotated dataset of paired human-written and GPT-3.5-generated science news articles from ScienceAlert, and shows that a Guided Few-shot prompting template with a single example boosts LLMs (including open-weight LLaMA-3 70b) to match or exceed graduate-stude
doi
10.1145/3720553.3746674
isbn
979-8-4007-1534-1
name
Can LLMs Beat Humans on Discerning Human-written and LLM-generated Science News?
source
pdf
acm_url
https://dl.acm.org/doi/10.1145/3720553.3746674
authors
Dominik Soós, Meng Jiang, Jian Wu
doi_url
https://doi.org/10.1145/3720553.3746674
license
CC BY 4.0
summary
The paper introduces SANews, a manually annotated dataset of paired human-written and GPT-3.5-generated science news articles from ScienceAlert, and shows that a Guided Few-shot prompting template with a single example boosts LLMs (including open-weight LLaMA-3 70b) to match or exceed graduate-stude
published
2025-09-15
conference
HT '25: 36th ACM Conference on Hypertext and Social Media, Chicago, IL, USA, September 15-19, 2025
acm_html_url
https://dl.acm.org/doi/full/10.1145/3720553.3746674
displayAuthor
Dominik Soós, Meng Jiang, Jian Wu
displayPublishTime
2025-09-15