Concept · 1 episode(s)

Political Bias in LLMs

← all concepts

Definition

Political bias in LLMs is the systematic tilt of model outputs on politically loaded questions — framings, refusals, recommendations. It’s sticky in ways that are partly data-driven and partly procedure-driven, and notoriously hard to measure without baking in your own framing.

Episodes covering this

Worth reading next

Papers we haven't done a deep dive on yet, but would recommend on this topic.