VerIF: Verification Engineering for Reinforcement Learning in Instruction Following
By: Hao Peng , Yunjia Qi , Xiaozhi Wang and more
Potential Business Impact:
Makes AI follow instructions better and more reliably.
Reinforcement learning with verifiable rewards (RLVR) has become a key technique for enhancing large language models (LLMs), with verification engineering playing a central role. However, best practices for RL in instruction following remain underexplored. In this work, we explore the verification challenge in RL for instruction following and propose VerIF, a verification method that combines rule-based code verification with LLM-based verification from a large reasoning model (e.g., QwQ-32B). To support this approach, we construct a high-quality instruction-following dataset, VerInstruct, containing approximately 22,000 instances with associated verification signals. We apply RL training with VerIF to two models, achieving significant improvements across several representative instruction-following benchmarks. The trained models reach state-of-the-art performance among models of comparable size and generalize well to unseen constraints. We further observe that their general capabilities remain unaffected, suggesting that RL with VerIF can be integrated into existing RL recipes to enhance overall model performance. We have released our datasets, codes, and models to facilitate future research at https://github.com/THU-KEG/VerIF.
Similar Papers
IFDECORATOR: Wrapping Instruction Following Reinforcement Learning with Verifiable Rewards
Computation and Language
Teaches AI to follow instructions better and avoid cheating.
IFDECORATOR: Wrapping Instruction Following Reinforcement Learning with Verifiable Rewards
Computation and Language
Teaches AI to follow instructions better and avoid cheating.
VERIRL: Boosting the LLM-based Verilog Code Generation via Reinforcement Learning
Machine Learning (CS)
Teaches computers to write hardware code better.