Large Reasoning Models Learn Better Alignment from Flawed Thinking
Paper
•
2510.00938
•
Published
•
58
RL
-
-
What Characterizes Effective Reasoning? Revisiting Length, Review, and Structure of CoT
Paper • 2509.19284 • Published • 22 -
Learning to Reason as Action Abstractions with Scalable Mid-Training RL
Paper • 2509.25810 • Published • 5 -
Agent Learning via Early Experience
Paper • 2510.08558 • Published • 270 -
Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
Paper • 2506.01939 • Published • 187 -
OpenThoughts: Data Recipes for Reasoning Models
Paper • 2506.04178 • Published • 49 -
Skywork-Reward-V2: Scaling Preference Data Curation via Human-AI Synergy
Paper • 2507.01352 • Published • 56 -
First Return, Entropy-Eliciting Explore
Paper • 2507.07017 • Published • 23 -
Part I: Tricks or Traps? A Deep Dive into RL for LLM Reasoning
Paper • 2508.08221 • Published • 50 -
Every Step Evolves: Scaling Reinforcement Learning for Trillion-Scale Thinking Model
Paper • 2510.18855 • Published • 71 -
Revealing the Power of Post-Training for Small Language Models via Knowledge Distillation
Paper • 2509.26497 • Published -
Every Activation Boosted: Scaling General Reasoner to 1 Trillion Open Language Foundation
Paper • 2510.22115 • Published • 83 -
Balanced Actor Initialization: Stable RLHF Training of Distillation-Based Reasoning Models
Paper • 2509.00309 • Published -
APRIL: Active Partial Rollouts in Reinforcement Learning to Tame Long-tail Generation
Paper • 2509.18521 • Published -
Bridging Offline and Online Reinforcement Learning for LLMs
Paper • 2506.21495 • Published • 3 -
On the Interplay of Pre-Training, Mid-Training, and RL on Reasoning Language Models
Paper • 2512.07783 • Published • 36 -
Stabilizing Reinforcement Learning with LLMs: Formulation and Practices
Paper • 2512.01374 • Published • 95 -
Reward Reasoning Model
Paper • 2505.14674 • Published • 37 -
Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models
Paper • 2512.13607 • Published • 27