Location May Be All You Need: Quantifying Position Bias in Academic Retrieval-Augmented Generation (RAG)
| dc.rights.license | CC-BY-NC-ND | |
| dc.contributor.advisor | Gatt, Albert | |
| dc.contributor.author | Spedener, Lesley | |
| dc.date.accessioned | 2025-10-31T00:01:47Z | |
| dc.date.available | 2025-10-31T00:01:47Z | |
| dc.date.issued | 2025 | |
| dc.identifier.uri | https://studenttheses.uu.nl/handle/20.500.12932/50630 | |
| dc.description.sponsorship | Utrecht University | |
| dc.language.iso | EN | |
| dc.subject | This thesis examines and quantifies the presence of position bias, i.e., LLMs' tendency to focus on certain regions of the input context irrespective of task relevance, across academic RAG tasks. | |
| dc.title | Location May Be All You Need: Quantifying Position Bias in Academic Retrieval-Augmented Generation (RAG) | |
| dc.type.content | Master Thesis | |
| dc.rights.accessrights | Open Access | |
| dc.subject.courseuu | Artificial Intelligence | |
| dc.thesis.id | 55057 |
