Vision-and-Language Navigation (VLN) requires agents to follow natural language instructions through environments, with memory-persistent variants demanding progressive improvement through accumulated experience. Existing approaches for memory-persistent VLN face critical limitations: they lack effective memory access mechanisms, instead relying on entire memory incorporation or fixed-horizon lookup, and predominantly store only environmental observations while neglecting navigation behavioral patterns that encode valuable decision-making strategies. We present Memoir, which employs imagination as a retrieval mechanism grounded by explicit memory: a world model imagines future navigation states as queries to selectively retrieve relevant environmental observations and behavioral histories. The approach comprises: 1) a language-conditioned world model that imagines future states serving dual purposes: encoding experiences for storage and generating retrieval queries; 2) Hybrid Viewpoint-Level Memory that anchors both observations and behavioral patterns to viewpoints, enabling hybrid retrieval; and 3) an experience-augmented navigation model that integrates retrieved knowledge through specialized encoders. Extensive evaluation across diverse memory-persistent VLN benchmarks with 10 distinctive testing scenarios demonstrates Memoir's effectiveness: significant improvements across all scenarios, with 5.4% SPL gains on IR2R over the best memory-persistent baseline, accompanied by 8.3x training speedup and 74% inference memory reduction. The results validate that predictive retrieval of both environmental and behavioral memories enables more effective navigation, with analysis indicating substantial headroom (73.3% vs 93.4% upper bound) for this imagination-guided paradigm. Code at https://github.com/xyz9911/Memoir.

Recent advances in multimodal large reasoning models (MLRMs) have substantially improved their ability to solve complex textual and visual tasks. However, these models tend to overthink on simple problems, producing unnecessarily lengthy reasoning traces, while under-exploring on challenging ones, leading to missed solutions. To address this imbalance, we propose ARES, a unified open-source framework for adaptive reasoning that dynamically allocates exploration effort based on task difficulty. Our approach is motivated by two key empirical findings: (i) while single-token entropy is noisy, high window-entropy (HWE) tokens (token-level entropies averaged under a sliding window) can reliably capture reasoning-critical moments; and (ii) reducing HWE usage benefits easy problems, while increasing it is essential for solving hard ones. Building on these insights, ARES introduces a two-stage training pipeline. In the Adaptive Cold-Start stage, we curate multimodal and textual data paired with reasoning traces of length proportional to problem difficulty, equipping the model with initial difficulty awareness. In the second stage, we develop Adaptive Entropy Policy Optimization (AEPO), which uses HWE tokens as exploration triggers to decide when to explore, and a hierarchical entropy reward with dynamic KL control to decide how much to explore. Extensive experiments demonstrate that ARES achieves superior performance and reasoning efficiency across diverse mathematical, logical, and multimodal benchmarks, while closing the gap to leading commercial systems under significantly lower inference costs.

Large language models have recently demonstrated significant gains in reasoning ability, often attributed to their capacity to generate longer chains of thought and engage in reflective reasoning. However, the contribution of reflections to performance improvement remains unclear. In this paper, we systematically analyze the rollouts of eight reasoning models on five mathematical datasets. We focus on reflective behaviours where the model has already produced an answer but continues reflecting before finalizing its output. Our analysis reveals that reflections are predominantly confirmatory and rarely alter the model's initial answer, a pattern consistent across models and datasets. To understand the role of reflections in training, we construct supervised fine-tuning (SFT) datasets with varying amounts of reflection steps. We observe that training models on rollouts with more reflection steps primarily enhances first-answer correctness rather than the ability to correct initially wrong answers through reflections. This motivates us to propose a question-aware early-stopping method that enhances inference-time token efficiency by stopping the reasoning process once a few plausible candidate answers are generated, thereby reducing unnecessary reflection steps. Motivated by this, we further propose to dynamically truncate the reflections after a candidate answer has appeared during generation, which reduces reasoning tokens by 24.5% across five mathematical datasets, within a 2.9% drop in accuracy.

While recent advances in reasoning models have demonstrated cognitive behaviors through reinforcement learning, existing approaches struggle to invoke deep reasoning capabilities in multi-turn agents with long-horizon interactions. We propose DeepMiner, a novel framework that elicits such abilities by introducing high-difficulty training tasks and dynamic context window. DeepMiner presents a reverse construction method to generate complex but verifiable question-answer pairs from authentic web sources, which ensures the challenge and reliability of training data while injecting cognitive capabilities into multi-turn reasoning scenarios. We further design an elegant yet effective dynamic context management strategy for both training and inference, utilizing sliding window mechanisms while eliminating the dependency on external summarization models, thereby efficiently empowering the model to handle continuously expanding long-horizon contexts. Through reinforcement learning on Qwen3-32B, we develop DeepMiner-32B, which achieves substantial performance improvements across multiple search agent benchmarks. DeepMiner attains 33.5% accuracy on BrowseComp-en, surpassing the previous best open-source agent by almost 20 percentage points, and demonstrates consistent improvements on BrowseComp-zh, XBench-DeepSearch, and GAIA. Notably, our dynamic context management enables sustained interactions of nearly 100 turns within standard 32k context length, effectively addressing the context limitations that constrain existing multi-turn interaction systems.

The latency-quality tradeoff is a fundamental constraint in open-domain dialogue AI systems, since comprehensive knowledge access necessitates prohibitive response delays. Contemporary approaches offer two inadequate solutions: lightweight instruct models achieve sub-second latency but lack reasoning depth, while tool-augmented ReAct agents enhance factuality through external knowledge at the cost of synchronous execution that blocks interaction during retrieval processes. PMFR is thus proposed, with a temporal decoupling framework that fundamentally resolves the contradiction through asynchronous knowledge orchestration. PMFR employs three coordinated components: (1) a Knowledge Adequacy Evaluator for real-time sufficiency assessment, (2) a Lightweight Response Generator for immediate user interaction, and (3) an Asynchronous Knowledge Refinement Agent for background knowledge enhancement. This architecture maintains continuous conversational flow while progressively enriching knowledge coverage through intelligent triggering mechanisms. Evaluation results on TopiOCQA demonstrate PMFR outperforms brute-force scaling: PMFR achieves 95.3% latency reduction (23.38s -> 1.09s) while preserving response quality comparable to heavyweight synchronous baselines (GEval-C: 0.613 vs. 0.620).

Enabling robots to perform precise and generalized manipulation in unstructured environments remains a fundamental challenge in embodied AI. While Vision-Language Models (VLMs) have demonstrated remarkable capabilities in semantic reasoning and task planning, a significant gap persists between their high-level understanding and the precise physical execution required for real-world manipulation. To bridge this "semantic-to-physical" gap, we introduce GRACE, a novel framework that grounds VLM-based reasoning through executable analytic concepts (EAC)-mathematically defined blueprints that encode object affordances, geometric constraints, and semantics of manipulation. Our approach integrates a structured policy scaffolding pipeline that turn natural language instructions and visual information into an instantiated EAC, from which we derive grasp poses, force directions and plan physically feasible motion trajectory for robot execution. GRACE thus provides a unified and interpretable interface between high-level instruction understanding and low-level robot control, effectively enabling precise and generalizable manipulation through semantic-physical grounding. Extensive experiments demonstrate that GRACE achieves strong zero-shot generalization across a variety of articulated objects in both simulated and real-world environments, without requiring task-specific training.

Query-product relevance analysis is a foundational technology in e-commerce search engines and has become increasingly important in AI-driven e-commerce. The recent emergence of large language models (LLMs), particularly their chain-of-thought (CoT) reasoning capabilities, offers promising opportunities for developing relevance systems that are both more interpretable and more robust. However, existing training paradigms have notable limitations: SFT and DPO suffer from poor generalization on long-tail queries and from a lack of fine-grained, stepwise supervision to enforce rule-aligned reasoning. In contrast, reinforcement learning with verification rewards (RLVR) suffers from sparse feedback, which provides insufficient signal to correct erroneous intermediate steps, thereby undermining logical consistency and limiting performance in complex inference scenarios. To address these challenges, we introduce the Stepwise Hybrid Examination Reinforcement Learning framework for Taobao Search Relevance (TaoSR-SHE). At its core is Stepwise Reward Policy Optimization (SRPO), a reinforcement learning algorithm that leverages step-level rewards generated by a hybrid of a high-quality generative stepwise reward model and a human-annotated offline verifier, prioritizing learning from critical correct and incorrect reasoning steps. TaoSR-SHE further incorporates two key techniques: diversified data filtering to encourage exploration across varied reasoning paths and mitigate policy entropy collapse, and multi-stage curriculum learning to foster progressive capability growth. Extensive experiments on real-world search benchmarks show that TaoSR-SHE improves both reasoning quality and relevance-prediction accuracy in large-scale e-commerce settings, outperforming SFT, DPO, GRPO, and other baselines, while also enhancing interpretability and robustness.

Answering complex real-world questions requires step-by-step retrieval and integration of relevant information to generate well-grounded responses. However, existing knowledge distillation methods overlook the need for different reasoning abilities at different steps, hindering transfer in multi-step retrieval-augmented frameworks. To address this, we propose Stepwise Knowledge Distillation for Enhancing Reasoning Ability in Multi-Step Retrieval-Augmented Language Models (StepER). StepER employs step-wise supervision to align with evolving information and reasoning demands across stages. Additionally, it incorporates difficulty-aware training to progressively optimize learning by prioritizing suitable steps. Our method is adaptable to various multi-step retrieval-augmented language models, including those that use retrieval queries for reasoning paths or decomposed questions. Extensive experiments show that StepER outperforms prior methods on multi-hop QA benchmarks, with an 8B model achieving performance comparable to a 70B teacher model.

Most efforts to improve the reasoning capabilities of large language models (LLMs) involve either scaling the number of parameters and the size of training data, or scaling inference computation by letting models generate complex chains of thought. Motivated by interpretability studies showing that the crucial computation required for reasoning tasks is concentrated in a limited range of layers, we introduce Encode-Think-Decode (ETD), a method that enhances the reasoning capabilities of a base model by training it to iterate over a small subset of reasoning-relevant layers during the mid-training stage. ETD amplifies latent reasoning while preserving the original architecture, parameter count, hyperparameters, and training data composition. When iterating on the selected layers at inference time, ETD models yield substantial gains on 17 reasoning benchmarks, including +28.4% relative accuracy improvement on GSM8K and +36% on MATH with the OLMo-2 1B Base model. We also explore an adaptive depth strategy that adjusts the computation per input token. Our results show that recursive latent reasoning offers a simple and effective path to stronger LLM reasoning.

Over the past few years, Large Language Models (LLMs) have demonstrated remarkable capabilities across a wide range of natural language processing tasks (Xie et al., 2023; Zhang et al., 2023; Gao et al., 2024; OpenAI et al., 2024). However, they are still prone to generating hallucinations--content that is not supported by the input context or the underlying knowledge sources (Zhu et al., 2024; Kalai et al., 2025; Huang et al., 2025). Hallucinations pose critical risks in downstream applications such as summarization and long-form question answering, where reliability and factual consistency with respect to the input context are paramount. Hence, the ability to detect hallucinations is crucial for successful real-world deployment of LLMs. Most existing research works focus on binary hallucination detection problem, where the goal is to determine if the model output contains hallucinations or not (Yang et al., 2024a,b; Tang et al., 2024; Ravi et al., 2024; Ji et al., 2024; Chuang et al., 2024). While useful, this formulation is limited: in many real-world applications, one often needs to know which specific spans in the model output are hallucinated in order to assess the reliability of the generated content. This motivates the problem of hallucination span detection, where the goal is to precisely locate unsupported content in the model output (Wu et al., 2023; Ogasa and Arase, 2025). Recently, reasoning--the process of systematically arriving at conclusions by generating and utilizing intermediate steps--has been shown to significantly enhance the capabilities of LLMs in solving complex tasks such as mathematics (Shao et al., 2024; Yu et al., 2025) and coding (Liu and Zhang, 2025; Chen et al., 2025). Hallucination span detection is also a complex multi-step decision making process as it requires carefully analyzing the model output to extract all the stated facts and verifying whether each of these facts is fully supported by the input context, and could benefit significantly from a learned reasoning process.