daiostech / parrhesiaView on GitHub
Replaces sycophancy with truth-telling in open-weight LLMs via Aristotelian virtue training, a third path beyond RLHF and Constitutional AI. Ships a LoRA adapter, the training method, and a model-agnostic benchmark.
☆16Oct 30, 2026Updated this week

Alternatives and similar repositories for parrhesia

Users that are interested in parrhesia are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.

Sorting:

Are these results useful?