Tag

GRPO

Every GRPO story we've curated in Bowl of Data, newest issue first — part of our weekly digest across AI, security, blockchain, and engineering.

4 items · 2 issues

AI & ML

Week 29 · 2026

Ring-Zero: Scaling Zero RL to a Trillion Parameters for Emergent Reasoning
This paper presents the development of Ring-2.5-1T-Zero, a trillion-parameter model trained via zero-shot reinforcement learning to elicit emergent reasoning. The study validates that massive scaling enables models to spontaneously develop complex problem-solving strategies like self-verification without human-annotated data.

AI & ML HuggingFace Papers Source ↗
Ring-Zero: Scaling Zero RL to a Trillion Parameters for Emergent Reasoning
This paper presents the development of Ring-2.5-1T-Zero, a trillion-parameter model trained via zero-shot reinforcement learning to elicit emergent reasoning. The study validates that massive scaling enables models to spontaneously develop complex problem-solving strategies like self-verification without human-annotated data.

AI & ML HuggingFace Papers Source ↗

Is One Layer Enough? Training A Single Transformer Layer Can Match Full-Parameter RL Training
This study demonstrates that reinforcement learning post-training for LLMs is not a uniform process across all parameters but is concentrated in specific middle layers. By leveraging this discovery, researchers developed layer-aware training methods that outperform traditional full-parameter optimization.

AI & ML arXiv Source ↗
Is One Layer Enough? Training A Single Transformer Layer Can Match Full-Parameter RL Training
This study demonstrates that reinforcement learning post-training for LLMs is not a uniform process across all parameters but is concentrated in specific middle layers. By leveraging this discovery, researchers developed layer-aware training methods that outperform traditional full-parameter optimization.

AI & ML arXiv Source ↗