Smaller, Weaker, Yet Better: Training LLM Reasoners via Compute-Optimal Sampling
Training on high-quality synthetic data from strong language models (LMs) is a common strategy to improve the reasoning performance of LMs. In this work, we revisit whether this strategy is compute-optimal under a fixed inference budget (e.g., FLOPs). To do so, we investigate the trade-offs between generating synthetic data using a stronger but more expensive (SE) model versus a weaker but cheaper (WC) model. We evaluate the generated data across three key metrics: coverage, diversity, and false positive rate, and show that the data from WC models may have higher coverage and diversity, but also exhibit higher false positive rates. We then finetune LMs on data from SE and WC models in different settings: knowledge distillation, self-improvement, and a novel weak-to-strong improvement setup where a weaker LM teaches reasoning to a stronger LM. Our findings reveal that models finetuned on WC-generated data consistently outperform those trained on SE-generated data across multiple benchmarks and multiple choices of WC and SE models. These results challenge the prevailing practice of relying on SE models for synthetic data generation, suggesting that WC may be the compute-optimal approach for training advanced LM reasoners.
Code (0)
등록된 구현이 없습니다.
Tasks
DiversityKnowledge DistillationSynthetic Data GenerationSimilar Papers 제목 키워드 기반
Putting the Value Back in RL: Better Test-Time Scaling by Unifying LLM Reasoners With Verifiers
Prevalent reinforcement learning~(RL) methods for fine-tuning LLM reasoners, such as GRPO or Leave-one-out PPO, abandon the learned value function in favor of empirically estimated returns. This hinders test-time compute…
MathReinforcement Learning (RL)How to train your ViT? Data, Augmentation, and Regularization in Vision Transformers
Vision Transformers (ViT) have been shown to attain highly competitive performance for a wide range of vision applications, such as image classification, object detection and semantic image segmentation. In comparison to…
Data Augmentationimage-classificationImage ClassificationImage Segmentation+4Small Models Struggle to Learn from Strong Reasoners
Large language models (LLMs) excel in complex reasoning tasks, and distilling their reasoning capabilities into smaller models has shown promise. However, we uncover an interesting phenomenon, which we term the Small Mod…
Incentivizing Reasoning from Weak Supervision
Large language models (LLMs) have demonstrated impressive performance on reasoning-intensive tasks, but enhancing their reasoning abilities typically relies on either reinforcement learning (RL) with verifiable signals o…
reinforcement-learningReinforcement LearningReinforcement Learning (RL)The Influence of Human-inspired Agentic Sophistication in LLM-driven Strategic Reasoners
The rapid rise of large language models (LLMs) has shifted artificial intelligence (AI) research toward agentic systems, motivating the use of weaker and more flexible notions of agency. However, this shift raises key qu…