Improving the mathematical reasoning capabilities of Large Language Models (LLMs) is critical for advancing artificial intelligence. However, access to extensive, diverse, and high-quality reasoning datasets remains a significant challenge, particularly for the open-source community. In this paper, we propose ScaleQuest, a novel, scalable, and cost-effective data synthesis method that enables the generation of large-scale mathematical reasoning datasets using lightweight 7B-scale models. ScaleQuest introduces a two-stage question-tuning process comprising Question Fine-Tuning (QFT) and Question Preference Optimization (QPO) to unlock the question generation capabilities of problem-solving models. By generating diverse questions from scratch – without relying on powerful proprietary models or seed data – we produce a dataset of 1 million problem-solution pairs. Our experiments demonstrate that models trained on our data outperform existing open-source datasets in both in-domain and out-of-domain evaluations. Furthermore, our approach shows continued performance improvement as the volume of training data increases, highlighting its potential for ongoing data scaling. The extensive improvements observed in code reasoning tasks demonstrate the generalization capabilities of our proposed method. Our work provides the open-source community with a practical solution to enhance the mathematical reasoning abilities of LLMs.

Generative Reward Models (GenRMs) leverage synthesized Chains of Thought (CoT) to reduce the need for massive labeled data, but this approach introduces risks of overoptimization due to the inability to guarantee the correctness of the CoTs. Identifying and optimizing unexpected behaviors within these synthesized CoT remains a challenge, as it heavily depends on precise annotations of intermediate behavior, similar to process supervision. In this work, we introduce a criteria-based preference tree for reward modeling, where each path in the tree represents a reasoning trajectory based on synthesized criteria. Crucially, each reasoning trajectory can be independently optimized through RL algorithm. These fine-grained process reward signals are derived from the inference-time computations and predefined rules, eliminating the need for human supervision. In experiments, SyncPL showed significant improvements over baselines on multiple human preference benchmarks. We further demonstrate that synthesized data can be learned using a long CoT format, analogous to an o1-like model, further enhancing performance while keeping stability and efficiency during training.

As large language models (LLMs) are increasingly applied to complex scientific problem-solving, their effectiveness is often limited by unconscious or failed tool usage. To address this issue, we introduce the Tool-Awareness Training (TAT) method, designed to enhance scientific reasoning. This approach leverages both forward and backward data generation strategies to strengthen the model’s conscious and selective tool utilization in multi-step reasoning tasks. Our method unfolds in three stages: (1) developing tool-knowledge through backward tooluse data generation (2) enhancing tool-awareness in multi-step reasoning by utilizing forward reasoning data, and (3) improving domain adaptability through large-scale domain-specific data for multi-task learning. These three stages progressively establish the foundation for tool learning and scientific reasoning, effectively integrating both, enabling the model to tackle multi-domain scientific tasks while optimizing tool usage. Our experimental results demonstrate that TAT significantly enhances LLM performance in mathematical and scientific reasoning tasks, particularly by improving the model’s tool utilization capabilities, including proactivity and execution success rates.

The OpenAI o1-series models have demonstrated that leveraging long-form Chain of Thought (CoT) can substantially enhance performance. However, the recursive thinking capabilities of Large Language Models (LLMs) remain limited, particularly in the absence of expert-curated data for distillation. In this paper, we propose AvR: Alignment via Refinement, a novel method aimed at unlocking the potential of LLMs for recursive reasoning through long-form CoT. AvR introduces a refinement process that integrates criticism and improvement actions, guided by differentiable learning techniques to optimize refinement-aware rewards. As a result, the synthesized multi-round data can be organized as a long refinement thought, further enabling test-time scaling. Experimental results show that AvR significantly outperforms conventional preference optimization methods. Notably, with only 3k synthetic samples, our method boosts the performance of the LLaMA-3-8B-Instruct model by over 20% in win rate on AlpacaEval 2.0. Our code is available at Github .

In inference-time scaling, Chain-of-Thought (CoT) plays a crucial role in enabling large language models (LLMs) to exhibit reasoning capabilities. However, in many scenarios, high-quality CoT data is scarce or even unavailable. In such cases, STaR-like methods can help LLMs synthesize CoT based on user queries and response, but they inevitably suffer from the risk of compounding errors. In this work, we tackle an even more challenging scenario: tool learning in the absence of user queries. We design a data scaling method using back-translation, which establishes an inference cycle to synthesize both user queries and CoT data. To reudce the compounding error of inference time, we introduce two rule-based verifiers to assess the validity of the synthesized CoT data. In particular, the Cycle Verifier facilitates performance improvement by continuously accumulating new data over multiple iterations. Our approach achieves a 75.4% pass rate and a 79.6% win rate using small models (7B) in StableToolBench. Notably, these results are obtained exclusively from self-synthesized high-quality data, without relying on external supervision or expert trajectories for warm-up.

Rather than merely to retain previously acquired generalization, achieving synergistic improvements between generalization and domain specialization in foundation models remains a significant challenge in both pre-training and post-training. As an alternative, we propose a test-time cross-domain knowledge integration method, Mixture of Multi-domain Agents (MMA), which dynamically combines the outputs of general-purpose and domain-specific models to enhance their performance on complex, domain‐specific tasks. MMA formulates the integration process as a search problem, using Monte Carlo Tree Search (MCTS) to find the path that optimally harmonizes the respective strengths of different models in generalization and domain-specific knowledge. In addition, We design specific action spaces to control the knowledge integration between multiple models, and cross-inspection reward is introduced to fairly score strategies in different domains. Experiments in diverse domains show that MMA can effectively combine the strengths of different models to enhance their performance. For instance, in legal tests, the average performance of all tasks increased from 42.57% to 53.68%. In financial tests, it improved from 56.01% to 62.68%.