Squeeze the Soaked Sponge: Efficient Off-policy Reinforcement Finetuning for Large Language Model
Reinforcement Learning (RL) has demonstrated its potential to improve the reasoning ability of Large Language Models (LLMs). One major limitation of most existing Reinforcement Finetuning (RFT) methods is that they are on-policy RL in nature, i.e., data generated during the past learning process is not fully utilized. This inevitably comes at a significant cost of compute and time, posing a stringent bottleneck on continuing economic and efficient scaling. To this end, we launch the renaissance of off-policy RL and propose Reincarnating Mix-policy Proximal Policy Gradient (ReMix), a general approach to enable on-policy RFT methods like PPO and GRPO to leverage off-policy data. ReMix consists of three major components: (1) Mix-policy proximal policy gradient with an increased Update-To-Data (UTD) ratio for efficient training; (2) KL-Convex policy constraint to balance the trade-off between stability and flexibility; (3) Policy reincarnation to achieve a seamless transition from efficient early-stage learning to steady asymptotic improvement. In our experiments, we train a series of ReMix models upon PPO, GRPO and 1.5B, 7B base models. ReMix shows an average Pass@1 accuracy of 52.10% (for 1.5B model) with 0.079M response rollouts, 350 training steps and achieves 63.27%/64.39% (for 7B model) with 0.007M/0.011M response rollouts, 50/75 training steps, on five math reasoning benchmarks (i.e., AIME'24, AMC'23, Minerva, OlympiadBench, and MATH500). Compared with 15 recent advanced models, ReMix shows SOTA-level performance with an over 30x to 450x reduction in training cost in terms of rollout data volume. In addition, we reveal insightful findings via multifaceted analysis, including the implicit preference for shorter responses due to the Whipping Effect of off-policy discrepancy, the collapse mode of self-reflection behavior under the presence of severe off-policyness, etc.
Code (0)
등록된 구현이 없습니다.
Tasks
Language ModelingLanguage ModellingLarge Language ModelMathReinforcement Learning (RL)Similar Papers 제목 키워드 기반
The SkipSponge Attack: Sponge Weight Poisoning of Deep Neural Networks
Sponge attacks aim to increase the energy consumption and computation time of neural networks. In this work, we present a novel sponge attack called SkipSponge. SkipSponge is the first sponge attack that is performed dir…
image-classificationImage ClassificationPolicy Finetuning: Bridging Sample-Efficient Offline and Online Reinforcement Learning
Recent theoretical work studies sample-efficient reinforcement learning (RL) extensively in two settings: learning interactively in the environment (online RL), or learning from an offline dataset (offline RL). However, …
Offline RLOpen-Ended Question Answeringreinforcement-learningReinforcement Learning+1Finetuning from Offline Reinforcement Learning: Challenges, Trade-offs and Practical Solutions
Offline reinforcement learning (RL) allows for the training of competent agents from offline datasets without any interaction with the environment. Online finetuning of such offline models can further improve performance…
DiversityOffline RLreinforcement-learningReinforcement Learning (RL)PROTO: Iterative Policy Regularized Offline-to-Online Reinforcement Learning
Offline-to-online reinforcement learning (RL), by combining the benefits of offline pretraining and online finetuning, promises enhanced sample efficiency and policy performance. However, existing methods, effective as t…
Computational Efficiencyreinforcement-learningReinforcement LearningReinforcement Learning (RL)Energy-Latency Attacks to On-Device Neural Networks via Sponge Poisoning
In recent years, on-device deep learning has gained attention as a means of developing affordable deep learning applications for mobile devices. However, on-device models are constrained by limited energy and computation…
Deep Learning