alt Hacker NewsThere's a large gap between:
1. This interpretation might not be "true" but it is a good-faith effort that respects the text.
2. This LLM is fabricating verses, chapters, and even books of the Bible.
If you've used LLMs much, you know that #2 is not only possible, its quite common. This is the kind of "might not be true" that you should be aware of when using an LLM for apologetics—or any effort where "truthiness" is important.
Replies
darrmit • 01/21/2025
I do agree point #2 is possible, but is that not something that could be accounted for and tuned in the model to an extent? I believe a tuned LLM would be able to distinguish between interpretation and generation of scripture, but I may be wrong.
➕ show 1 reply
While I kinda agree, many denominations are also questionable in the "good-faith effort that respects the text" department. Some like making up new chapters while others overemphasize specific short sections while disregarding many others.
Much more so than any decent LLM ever would. Even while they're "making stuff up", they largely stick to the general themes.