Awesome-AgenticLLM-RL-Papers
The companion repository for a survey paper on agentic reinforcement learning for large language models, tabulating RL algorithms such as PPO, DPO, GRPO and their variants with objective type, mechanism and reward signal.
This repository is the official companion to the survey paper 'The Landscape of Agentic Reinforcement Learning for LLMs.' It links the arXiv paper and its Hugging Face listing, and provides a citation entry.
Its main content is a structured table of reinforcement-learning methods used for training agentic LLMs, grouped into families such as PPO, DPO and GRPO. For each method it records the year, objective type, whether it clips the policy ratio, whether and how it applies a KL penalty, its key mechanism, the training signal it uses, and links to the paper and any released code or model.
It is organized as a reference table rather than narrative text.