Researchers debate RLVR's lineage: 'RLHF was the weird one, STaR was already around'
soldni · x · 2026-10-15
In a discussion on whether RLVR (RL with verifiable rewards) was a genuinely new idea, soldni argues the idea existed but known recipes weren't broadly available in the literature the way tool use recipes were.
lateinteraction counters that RLVR is just RL—RLHF is the more elaborate/weird one—and STaR arrived soon after; using RL for chain of thought in early 2021 would have been remarkable, but this wasn't that.
Related event: Researchers Trace RLVR's Roots: It's Just RL, and the Idea Isn't New(3 posts)→
More from Research
- DecisionRobust benchmark stress-tests 30 decision models against prompt injection and perturbations — Yifan Lan · 2026-10-15
- Continuum Attention: continuous integration beats softmax attention in all paired comparisons — Afshin Khadangi · 2026-10-15
- Researchers accuse an AI lab paper of overlapping their ARTS work without attribution — xwang_lk · 2026-10-15
- RootPath launches BinderVerse: AI-designed binders get wet-lab data in as few as 3 days — iskander · 2026-10-15
- New arXiv paper shows unconstrained MLIPs face exponentially ill-conditioned sampling limits — CatAstro_Piyush · 2026-10-15
- Paper: method-transferable ML potentials can outperform the quantum-chemical methods they trained on — CatAstro_Piyush · 2026-10-15