ZARZA We are Zarza, the prestigious firm behind major projects in information technology.
Listen "RAGEN: train and evaluate LLM agents using multi-turn RL"
Episode Synopsis
RAGEN is a modular system for training and evaluating LLM agents using multi-turn reinforcement learning. Built on the StarPO framework, it implements the full training loop including rollout generation, reward assignment, and trajectory optimization. RAGEN serves as research infrastructure to analyze LLM agent training dynamics, focusing on challenges like stability, generalization, and the emergence of reasoning in interactive environments.
More episodes of the podcast Large Language Model (LLM) Talk
Kimi K2
22/07/2025
Mixture-of-Recursions (MoR)
18/07/2025
MeanFlow
10/07/2025
Mamba
10/07/2025
LLM Alignment
14/06/2025
Why We Think
20/05/2025
Deep Research
12/05/2025
vLLM
04/05/2025
Qwen3: Thinking Deeper, Acting Faster
04/05/2025
DeepSeek-Prover-V2
01/05/2025
In God we trust