PardisSzah/FaithConflict

FaithConflict A controlled benchmark for alignment-induced unfaithfulness (AIU): the silent, undisclosed suppression or inversion of source content by aligned language models. FaithConflict is the dataset introduced in Emergent Unfaithfulness: How Alignment Training Causes Language Models to Silently Override Task Faithfulness (Zahraei, Singh, Tur, and Hakkani-Tur, COLM 2026). It isolates a single experimental question: if a document is held constant and only the truth value of… See the full description on the dataset page: https://huggingface.co/datasets/PardisSzah/FaithConflict.

<iframe src="https://silytics.si/embed/item/195031e6-4491-439a-ab6c-d76939cda18d" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 19 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.