belief ↔ Large Language Models

Relations (1)

related 0.30 — supporting 3 facts

The relationship is established through the academic study of how belief is attributed to and measured in large language models, as discussed by Herrmann and Levinstein in [1], [2], and [3].

Facts (3)

Sources

https://scholar.google.com/citations?view_op=view_... scholar.google.com Daniel A Herrmann, Benjamin A Levinstein · Springer Netherlands 3 facts

perspectiveDaniel A. Herrmann and Benjamin A. Levinstein argue that while measuring belief in large language models shares features with belief measurement in decision theory and formal epistemology, there are differences that necessitate changes in how belief is measured in large language models.

claimDaniel A. Herrmann and Benjamin A. Levinstein established four criteria for measuring belief in large language models, drawing from insights in philosophy and machine learning practices.

claimDaniel A. Herrmann and Benjamin A. Levinstein argue that the current field of studying belief in large language models lacks a unified theoretical foundation.