Small Reward Models via Backward Inference https://arxiv.org/abs/2602.13551…

Operationalising the Superficial Alignment Hypothesis via Task Complexity https://arxiv.org/abs/2602.15829 https://www.alphaxiv.org/ru/overview/2602.15829

@tokareviv мне кажется надо отдельный блок vibecoding создавать

Так других кодингов нет

RL Excursions during Pretraining: How early is too early for On-policy Learning? https://rl-excursions.github.io/

https://github.com/Shekswess/tiny-reasoning-language-model