MagicBeanStalk: Escaping the RLVR Zero Advantage Trap with Recursive Data Synthesis
cs.AI, cs.LG
Submitted: 2026-10-08
Updated: 2026-10-08
Sources
- Long Grounded Thoughts: Synthesizing Visual Problems and Reasoning Chains at Scale
- Big-Math: A Large-Scale, High-Quality Math Dataset for Reinforcement Learning in Language Models
- AceReason-Nemotron: Advancing Math and Code Reasoning through Reinforcement Learning
- Process Reinforcement through Implicit Rewards
- Beyond Benchmarks: MathArena as an Evaluation Platform for Mathematics with LLMs
- Omni-MATH: A Universal Olympiad Level Mathematic Benchmark For Large Language Models
- Skywork Open Reasoner 1 Technical Report
- BroRL: Scaling Reinforcement Learning via Broadened Exploration
- FinanceBench: A New Benchmark for Financial Question Answering
- LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
- DeltaPrompts: Escaping the Zero-Delta Trap in Multimodal Distillation
- Reasoning Core: A Scalable RL Environment for LLM Symbolic Reasoning
- SwS: Self-aware Weakness-driven Problem Synthesis in Reinforcement Learning for LLM Reasoning
- LongPerceptualThoughts: Distilling System-2 Reasoning for System-1 Perception
- rStar-Coder: Scaling Competitive Code Reasoning with a Large-Scale Verified Dataset
- SCP-116K: A High-Quality Problem-Solution Dataset and a Generalized Pipeline for Automated Extraction in the Higher Education Science Domain
- Retro-Search: Exploring Untaken Paths for Deeper and Efficient Reasoning
- Golden Goose: A Simple Trick to Synthesize Unlimited RLVR Tasks from Unverifiable Internet Text
- POPE: Learning to Reason on Hard Problems via Privileged On-Policy Exploration
- REASONING GYM: Reasoning Environments for Reinforcement Learning with Verifiable Rewards
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems