Method
PPO
Proximal policy optimization.
6 cases from 4 companies apply PPO. 2 were checked in an online A/B test.
Typical tasks: Label data with LLMs, Build an assistant over internal docs, Recommend from what the user just did, Optimize decisions. Companies: Rabota.ru, Netflix, SberDevices and 1 more. Often combined with: Reward model, RLHF, Supervised fine-tuning, DPO / preference optimization, Synthetic data generation (LLM) and 18 more.
| Post-Training Generative Recommenders with Advantage-Weighted Supervised Finetuning | media-streaming | 2025-10 | |
| Advanced Use of GigaChat at Rabota.ru: Synthetic Data and Search | classifieds | 2024-05 | |
| Well Placement for Drilling Using Reinforcement LearningDT | DT | manufacturing | 2024-05 |
| A Job Posting Summary Generation Service Based on an LM Fine-Tuned with RLHF | classifieds | 2023-05 | |
RLHF for retrieval based NLP system![]() | ![]() | general | 2023-05 |
| LLaMA and a Zillion Ways to Fine-Tune It | — | general | 2023-05 |
