Papers by Antonie Jetter
Analyzing the Dialect Diversity in Multi-document Summaries (2022.coling-1)
Copied to clipboard
| Challenge: | Social media posts provide compelling, yet challenging source of data of diverse perspectives . automatic text summarization algorithms preserve salient information from the source text . |
| Approach: | They propose to compress large collections of documents into short summaries that preserve salient information from the source text. |
| Outcome: | The proposed dataset shows that humans annotate fairly well-balanced dialect diverse summaries without loss of diversity. |
When Context Leads but Parametric Memory Follows in Large Language Models (2024.emnlp-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) have demonstrated remarkable progress in leveraging diverse knowledge sources. |
| Approach: | They propose to analyze how large language models prioritize and utilize local context and parametric knowledge when answering open-ended questions in knowledge-consistent scenarios. |
| Outcome: | The proposed model prioritizes parametric and contextual knowledge over parametric knowledge in knowledge-consistent scenarios and their tendency to hallucinate under varying context sizes. |