Topic · 28 episodes across 6 reviews

Training and Reinforcement Learning for LLM Agents

← all reviews

A cluster of papers probed what RL actually does to agents — when it teaches genuinely new skills, how it silently fails, how a model could sabotage it, and how broken baselines distorted a whole subfield.

Covered in these reviews

Related concepts

Episodes and papers for this topic live on the concept pages.

Training Methods Evaluation & Benchmarks Agentic AI GRPO Agentic RL Supervised Fine-Tuning Synthetic Data AI Alignment Rollout Sampling Reinforcement Learning Reward Model Trajectory Analysis Reward Hacking RL Post-Training