Sources
Nathan Lambert
OpenAI announced a new RL finetuning API. You can do this on your own models with Open Instruct -- the repo we used to train Tulu 3. Expanding reinforcement learning with verifiable rewards (RLVR) to more domains and with better answer extraction (what OpenAI calls a grader, a… https://t.co/VEBdH8AR28
Omar Khattab OpenAI is finally playing the specialization game now. Depending on how they do it, could be a very wise (certainly non-AGI-flavored!) decision. Late alignment (to a task and domain) is all you need? https://t.co/XRbON890QQ
Tibor Blaho OpenAI announced an expanded Reinforcement Fine-Tuning Research Program, allowing developers to customize AI models for domain-specific tasks by training them on datasets ranging from dozens to thousands of high-quality tasks and evaluating responses against reference answers -… https://t.co/zhtKPAtjzE
