To Mix or To Merge: Toward Multi-Domain Reinforcement Learning for Large Language Models

arXiv:2602.12566 · cs.AI · Submitted 2026-02-13 · Read on arXiv

cs.AI

Submitted: 2026-02-13

Updated: 2026-09-07

Comments: Accepted by COLM 2026

Code: https://github.com/Mosi-AI/M2RL

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Reinforcement Learning — 41× in this paper · explained in 4 episodes
• Verification — 31× in this paper · explained in 2 episodes
• Model Merging — 19× in this paper · explained in 2 episodes
• Distillation — 18× in this paper · explained in 2 episodes
• Benchmark — 12× in this paper · explained in 2 episodes
• On-Policy Distillation — 10× in this paper · explained in 3 episodes
• Verifiable Rewards — 8× in this paper · explained in 2 episodes
• Accuracy — 7× in this paper · explained in 1 episode
• Adam — 7× in this paper · explained in 2 episodes
• Multi-Teacher On-Policy Distillation — 7× in this paper
• Cosine Similarity — 6× in this paper · explained in 3 episodes
• Fine-Tuning — 6× in this paper · explained in 5 episodes
• Reinforcement Learning with Verifiable Rewards — 6× in this paper
• Robustness — 6× in this paper · explained in 5 episodes
• Supervised Fine-Tuning — 6× in this paper
• Consistency — 5× in this paper · explained in 3 episodes
• Verifier — 4× in this paper · explained in 2 episodes
• Trace — 3× in this paper · explained in 2 episodes
• GRPO — 2× in this paper · explained in 6 episodes
• Large Language Models (LLMs) — 2× in this paper · explained in 6 episodes
• Thinking mode — 2× in this paper · explained in 2 episodes
• Ablation — 1× in this paper · explained in 1 episode
• Ablation study — 1× in this paper · explained in 1 episode
• Agentic Reasoning — 1× in this paper · explained in 1 episode
• Alignment — 1× in this paper · explained in 1 episode
• Chain-of-Thought (CoT) — 1× in this paper · explained in 4 episodes
• Chain-of-Thought (CoT) Reasoning — 1× in this paper · explained in 2 episodes
• Core — 1× in this paper · explained in 2 episodes
• Curse of Dimensionality — 1× in this paper · explained in 6 episodes
• Deep Reinforcement Learning — 1× in this paper · explained in 2 episodes
• Fisher Information — 1× in this paper · explained in 3 episodes
• Foundation Model — 1× in this paper · explained in 3 episodes
• Foundation Models — 1× in this paper · explained in 7 episodes
• Group Relative Policy Optimization — 1× in this paper
• Group Relative Policy Optimization (GRPO) — 1× in this paper · explained in 5 episodes
• Kullback-Leibler (KL) Divergence — 1× in this paper · explained in 3 episodes
• Pearson Correlation Coefficient — 1× in this paper
• Pruning — 1× in this paper · explained in 4 episodes
• Reasoning Traces — 1× in this paper · explained in 3 episodes
• Reinforcement Learning (RL) — 1× in this paper · explained in 24 episodes
• Reward Function — 1× in this paper · explained in 2 episodes
• Reward Hacking — 1× in this paper · explained in 8 episodes
• Supervised Fine-Tuning (SFT) — 1× in this paper · explained in 8 episodes
• Transformer — 1× in this paper · explained in 2 episodes

Sources

• Big-Math: A Large-Scale, High-Quality Math Dataset for Reinforcement Learning in Language Models
• Nemotron 3 Nano: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning
• FullStack Bench: Evaluating LLMs as Full Stack Coders
• R1-Code-Interpreter: LLMs Reason with Code via Supervised and Multi-stage Reinforcement Learning
• Process Reinforcement through Implicit Rewards
• Agentic Reinforced Policy Optimization
• Soft Adaptive Policy Optimization
• DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
• Skywork Open Reasoner 1 Technical Report
• A Systematic Study of In-the-Wild Model Merging for Large Language Models
• Qwen2.5-Coder Technical Report
• Editing Models with Task Arithmetic
• OpenAI o1 System Card
• LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
• Adam: A Method for Stochastic Optimization
• Tulu 3: Pushing Frontiers in Open Language Model Post-Training
• Competition-Level Code Generation with AlphaCode
• Humanity's Last Exam
• Generalizing Verifiable Instruction Following
• DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models

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
• 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
• VSAL: A Vision Solver with Adaptive Layouts for Graph Property Detection