Bibo Cai

Yang He, Xiao Ding, Bibo Cai et al.

Tool-Integrated Reasoning (TIR) extends LLM capabilities by leveraging external environments. However, existing methods lack the deliberation during sequential tool invocation required for strategic planning and self-correction. While RL mitigates this, conventional approaches for Tool-Integrated Reasoning are hindered by sparse outcome-based rewards, failing to supervise intermediate reasoning steps and tool invocations. To address this, we propose DeepTool, a novel framework that scales deliberate thinking within the interleaved process of thinking, action, and observation at each turn. In DeepTool, we first introduce a synthesis pipeline that evolves extended thinking into interleaved trajectories, integrating adversarial perturbations to ensure robustness and self-correction. Secondly, we devise Process-Supervised Reinforcement Learning based on GRPO, which utilizes an Action-Centric Process Reward to reinforce intermediate interleaved thinking and enforce precise tool invocation at every turn. Extensive experiments demonstrate that DeepTool achieves superior performance, boosting Qwen2.5-7B significantly across six benchmarks (e.g., AIME24: 3.2% -> 40.4% and HMMT25: 0.0% -> 28.6%). Furthermore, the token cost-effectiveness analysis confirms the utility of interleaved thinking, demonstrating DeepTool's optimal balance between performance and token efficiency.

Zhouhao Sun, Xuan Zhang, Xiao Ding et al.

Currently, process reward models (PRMs) have exhibited remarkable potential for test-time scaling. Since large language models (LLMs) regularly generate flawed intermediate reasoning steps when tackling a broad spectrum of reasoning and decision-making tasks, PRMs are required to possess capabilities for detecting process-level errors in real-world scenarios. However, existing benchmarks primarily focus on mathematical reasoning, thereby failing to comprehensively evaluate the error detection ability of PRMs across diverse reasoning scenarios. To mitigate this gap, we introduce GR-Ben, a process-level benchmark specifically designed for assessing PRM's performance across two primary reasoning domains (science and logic) and nine subdomains. We conduct extensive experiments on a diverse set of 22 models, encompassing both PRMs and LLMs, and derive two key findings: (1) In domains beyond mathematical reasoning, the error-detection ability of existing PRMs and LLMs is found to be markedly weaker by comparison.(2) In general, PRMs are less adept at identifying knowledge-based errors, whereas LLMs exhibit poorer performance in detecting computational errors.We hope GR-Ben can foster future researches on PRMs for general domains, thereby enhancing the reasoning capabilities of LLMs.

Yirong Zeng, Yufei Liu, Xiao Ding et al.

A central belief in scaling reinforcement learning with verifiable rewards for instruction following (IF) tasks is that, a diverse mixture of verifiable hard and unverifiable soft constraints is essential for generalizing to unseen instructions. In this work, we challenge this prevailing consensus through a systematic empirical investigation. Counter-intuitively, we find that models trained on hard-only constraints consistently outperform those trained on mixed datasets. Extensive experiments reveal that reward precision, rather than constraint diversity, is the primary driver of effective alignment. The LLM judge suffers from a low recall rate in detecting false response, which leads to severe reward hacking, thereby undermining the benefits of diversity. Furthermore, analysis of the attention mechanism reveals that high-precision rewards develop a transferable meta-skill for IF. Motivated by these insights, we propose a simple yet effective data-centric refinement strategy that prioritizes reward precision. Evaluated on five benchmarks, our approach outperforms competitive baselines by 13.4\% in performance while achieving a 58\% reduction in training time, maintaining strong generalization beyond instruction following. Our findings advocate for a paradigm shift: moving away from the indiscriminate pursuit of data diversity toward high-precision rewards.

Yang Zhao, Yangou Ouyang, Xiao Ding et al.

While Hybrid Supervised Fine-Tuning (SFT) followed by Reinforcement Learning (RL) has become the standard paradigm for training LLM agents, effective mechanisms for data allocation between these stages remain largely underexplored. Current data arbitration strategies often rely on surface-level heuristics that fail to diagnose intrinsic learning needs. Since SFT targets pattern consolidation through imitation while RL drives structural adaptation via exploration, misaligning data with these functional roles causes severe optimization interference. We propose PRISM, a dynamics-aware framework grounded in Schema Theory that arbitrates data based on its degree of cognitive conflict with the model's existing knowledge. By analyzing the spatial geometric structure of gradients, PRISM identifies data triggering high spatial concentration as high-conflict signals that require RL for structural restructuring. In contrast, data yielding diffuse updates is routed to SFT for efficient consolidation. Extensive experiments on WebShop and ALFWorld demonstrate that PRISM achieves a Pareto improvement, outperforming state-of-the-art hybrid methods while reducing computational costs by up to 3.22$\times$. Our findings suggest that disentangling data based on internal optimization regimes is crucial for scalable and robust agent alignment.

Yang Zhao, Hepeng Wang, Xiao Ding et al.

Group-Relative Policy Optimization (GRPO) has emerged as an efficient paradigm for aligning Large Language Models (LLMs), yet its efficacy is primarily confined to domains with verifiable ground truths. Extending GRPO to open-domain settings remains a critical challenge, as unconstrained generation entails multi-faceted and often conflicting objectives - such as creativity versus factuality - where rigid, static reward scalarization is inherently suboptimal. To address this, we propose MAESTRO (Meta-learning Adaptive Estimation of Scalarization Trade-offs for Reward Optimization), which introduces a meta-cognitive orchestration layer that treats reward scalarization as a dynamic latent policy, leveraging the model's terminal hidden states as a semantic bottleneck to perceive task-specific priorities. We formulate this as a contextual bandit problem within a bi-level optimization framework, where a lightweight Conductor network co-evolves with the policy by utilizing group-relative advantages as a meta-reward signal. Across seven benchmarks, MAESTRO consistently outperforms single-reward and static multi-objective baselines, while preserving the efficiency advantages of GRPO, and in some settings even reducing redundant generation.

Yirong Zeng, Shen You, Yufei Liu et al.

Equipping LLMs with external tools effectively addresses internal reasoning limitations. However, it introduces a critical yet under-explored phenomenon: tool overuse, the unnecessary tool-use during reasoning. In this paper, we first reveal this phenomenon is pervasive across diverse LLMs. We then experimentally elucidate its underlying mechanisms through two key lenses: (1) First, by analyzing tool-use behavior across different internal knowledge availability regions, we identify a \textit{knowledge epistemic illusion}: models misjudge internal knowledge boundaries and fail to accurately perceive their actual knowledge availability. To mitigate this, we propose a knowledge-aware epistemic boundary alignment strategy based on direct preference optimization, which reduces tool usage in by 82.8\% while yielding an accuracy improvement. (2) Second, we establish a causal link between reward structures and tool-use behavior by visualizing the tool-augmented training process. It reveals that \textit{outcome-only rewards} inadvertently encourage tool overuse by rewarding only final correctness, regardless of tool efficiency. To verify this, we balance reward signals during training rather than relying on outcome-only rewards, cutting unnecessary tool calls by 66.7\% (7B) and 60.7\% (32B) without sacrificing accuracy. Finally, we provide theoretical justification in this two lenses to understand tool overuse.

Yang He, Xiao Ding, Bibo Cai et al.

While reasoning-augmented large language models (RLLMs) significantly enhance complex task performance through extended reasoning chains, they inevitably introduce substantial unnecessary token consumption, particularly for simpler problems where Short Chain-of-Thought (Short CoT) suffices. This overthinking phenomenon leads to inefficient resource usage without proportional accuracy gains. To address this issue, we propose Self-Route, a dynamic reasoning framework that automatically selects between general and reasoning modes based on model capability estimation. Our approach introduces a lightweight pre-inference stage to extract capability-aware embeddings from hidden layer representations, enabling real-time evaluation of the model's ability to solve problems. We further construct Gradient-10K, a model difficulty estimation-based dataset with dense complexity sampling, to train the router for precise capability boundary detection. Extensive experiments demonstrate that Self-Route achieves comparable accuracy to reasoning models while reducing token consumption by 30-55\% across diverse benchmarks. The proposed framework demonstrates consistent effectiveness across models with different parameter scales and reasoning paradigms, highlighting its general applicability and practical value.

Jinglong Gao, Xiao Ding, Lingxiao Zou et al.

Recent studies provide large language models (LLMs) with textual task-solving experiences via prompts to improve their performance. However, previous methods rely on substantial human labor or time to gather such experiences for each task, which is impractical given the growing variety of task types in user queries to LLMs. To address this issue, we design an autonomous experience transfer framework to explore whether LLMs can mimic human cognitive intelligence to autonomously transfer experience from existing source tasks to newly encountered target tasks. This not only allows the acquisition of experience without extensive costs of previous methods, but also offers a novel path for the generalization of LLMs. Experimental results on 13 datasets demonstrate that our framework effectively improves the performance of LLMs. Furthermore, we provide a detailed analysis of each module in the framework.

Zhouhao Sun, Xiao Ding, Li Du et al.

Large language models (LLMs) have achieved significant performance in various natural language reasoning tasks. However, they still struggle with performing first-order logic reasoning over formal logical theories expressed in natural language. This is because the previous LLMs-based reasoning systems have the theoretical incompleteness issue. As a result, it can only address a limited set of simple reasoning problems, which significantly decreases their generalization ability. To address this issue, we propose a novel framework, named Generalizable and Faithful Reasoner (GFaiR), which introduces the paradigm of resolution refutation. Resolution refutation has the capability to solve all first-order logic reasoning problems by extending reasoning rules and employing the principle of proof by contradiction, so our system's completeness can be improved by introducing resolution refutation. Experimental results demonstrate that our system outperforms previous works by achieving state-of-the-art performances in complex scenarios while maintaining performances in simple scenarios. Besides, we observe that GFaiR is faithful to its reasoning process.