Reddit is using LLMs to solve a problem LLMs largely created

Reddit is using large language models to detect and remove spam and policy-violating content from its platform. The company stated in a blog post that it deployed LLM-based detection tools to identify violations at scale. Reddit's previous spam-fighting systems relied on rule-based filters and user reports. The platform did not specify what percentage of spam the new system catches or how it compares to prior methods.

Next story