Year: 2026
Type: dissertation
Abstract: Large Language Models (LLMs) usually fall into the problem of hallucination and lack of reliable references for their generated answers. The primary cause of these limitations is the context provided for the model's input, and the lack of robustness of the LLM. For this reason, in this thesis, I explore the efficacy of the existing attribution methods for large language models, as well as their failure conditions.
Source: Aaltodoc (Aalto University)
Author: Trinh, Quoc
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7204018950
