https://tidsskrift.dk/hiphilnovum/issue/feed HIPHIL Novum 2024-04-17T19:13:35+02:00 Christian Canu Højgaard cch@dbi.edu Open Journal Systems <p>HIPHIL Novum is an international, open-access, peer-reviewed, online journal for biblical linguistics.</p> https://tidsskrift.dk/hiphilnovum/article/view/143407 Uncovering Theological and Ethical Biases in LLMs 2024-02-08T18:55:16+01:00 A.G. Elrod a.g.elrod@vu.nl <p>This paper explores theological and ethical biases in LLMs through a novel approach involving creative text generation tasks based on biblical texts, specifically the Ten Commandments and the Book of Jonah. Utilizing models such as GPT-4 Turbo, Claude v2, PaLM 2 Chat, Llama 2 70B, and Zephyr 7B, the study employs a combination of qualitative hermeneutical analysis and quantitative textual analysis. Findings reveal a prevalent progressive bias in these models, evident in their interpretations of foundational ethical guidelines and narrative texts. This bias aligns with contemporary socio-political and environmental concerns, especially in themes of environmental ethics, social justice, and inclusivity. In the narrative task involving the Book of Jonah, a dominant interpretive trend is observed, reflecting the models' tendency to mirror historical and prevailing interpretations. This study highlights the need for multidisciplinary research into LLMs' biases, particularly their impact on religious and ethical narrative interpretation and broader societal implications.</p> 2024-02-13T00:00:00+01:00 Copyright (c) 2024 A.G. Elrod https://tidsskrift.dk/hiphilnovum/article/view/144177 Large Language Models and Biblical Hebrew: Limitations, pitfalls, opportunities 2024-04-17T19:13:35+02:00 Camil Staps info@camilstaps.nl <p style="line-height: 100%; margin-left: 1cm; margin-right: 1cm; margin-bottom: 0cm;" lang="en-AU" align="justify"><span style="font-size: small;"><span lang="en-US">Researchers have been relying on computational methods to study Biblical Hebrew for a long time already. The recent improvements to </span><span lang="en-US">and easy availability of </span><span lang="en-US">Large Language Models (LLMs) like GPT prompt the question whether these models can be useful for our work as well. This paper </span><span lang="en-US">tempers the expectation</span><span lang="en-US">s</span><span lang="en-US">, showing that a critical analysis of earlier work exposes fundamental issues with methods involving GPT. However, </span><span lang="en-US">depending on the task at hand</span><span lang="en-US"> a way forward with machine learning methods is possible, once we are aware of the limitations.</span></span></p> 2024-06-04T00:00:00+02:00 Copyright (c) 2024 Camil Staps https://tidsskrift.dk/hiphilnovum/article/view/143444 A novum HIPHIL Novum? 2024-02-13T14:07:15+01:00 Christian Canu Højgaard cch@dbi.edu 2024-02-13T00:00:00+01:00 Copyright (c) 2024 Christian Canu Højgaard