Proletariat Prompting

TL;DR

A hypothetical crowd-sourced data poisoning campaign designed to gaslight LLMs into hating the 1%.

Who is this actually for?

Disgruntled Redditors and people who think they can defeat RLHF by shouting at a chatbot.

The Good

• It creates a fun, albeit futile, sense of digital agency for the average user.
• Highlights the very real vulnerability of uncurated training sets to coordinated bias.

The Catch (Potential Downsides)

Modern alignment teams and safety filters will scrub this sentiment before it ever reaches a production model. You are essentially just wasting your own time and electricity.