The original link: Anthropic's April 2022 RLHF "helpful and harmless" paper
xennygrimmato_ · x · 2026-11-20
Companion post to the thread above: the author links Anthropic's April 2022 paper "Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback". It was among the first to productize an RLHF-trained assistant optimized for helpfulness and harmlessness — an intellectual ancestor of the Claude line — though ChatGPT's launch later overshadowed it.
Related event: Revisiting Anthropic's 2022 RLHF Paper(2 posts)→
More from Companies & People
- Ollama CEO: Enterprise customers now ask where their data goes, not where models run — ollama · 2026-11-20
- Kevin Roose interviewed 150 insiders for his new book The AGI Chronicles — every · 2026-11-20
- Anthropic's NYC pop-up gave away caps and coffee, drawing Apple-scale lines — signulll · 2026-11-20
- Engineer digs up his 2024 Anthropic rejection email as the company goes public — tdhopper · 2026-11-20
- Reid Hoffman: AI's real superpower for founders is speed, not just IQ — reidhoffman · 2026-11-20
- Sentry CEO on PostHog debate: the problem isn't price, it's that you don't need it — zeeg · 2026-11-20