Multi-chapter novel generation requires models to balance two quality dimensions simultaneously: chapter-level reader engagement and global narrative rhythm. Existing approaches have mainly focused on planning, memory, and execution, while paying limited attention to the mismatch between training signals and reader experience. We propose **ChapterRL**, which reformulates multi-chapter generation as a reward design problem. Specifically, ChapterRL introduces two complementary reward signals. **Reading Depth Residual** (RDR) derives debiased chapter-level quality signals from reader continuation and completion behavior through genre–position stratified normalization. **Dynamic Rhythm Trajectory** (DRT) uses an arousal estimator to impose a threshold-gated penalty on systematic rhythm collapse in the later part of a story. The two signals correspond to chapter attractiveness and global narrative momentum, and are jointly integrated into reinforcement learning training. We evaluate the full ChapterRL policy on three automatic benchmarks and separately validate its RDR reward model on task-aligned narrative domains of the Chinese WritingPreferenceBench. Results show that ChapterRL improves chapter quality and rhythm stability over supervised fine-tuning and single-reward variants, while RDR exhibits strong narrative preference alignment as an efficient scalar reward model.Code is available at [ChapterRL](https://github.com/SorBetQ/ChapterRL).

Abundant procedural knowledge on the Web holds great potential for helping agents solve long-horizon tasks. However, such knowledge is often multimodal, heterogeneous, noisy, and implicitly assumes human executors, making it difficult to use directly as the skills required by agents. To bridge the gap between human-oriented guides and agent-executable skills, we formalize this problem as guide-to-skill learning: converting in-the-wild guides into executable skills and continuously improving them from trajectories observable to the agent. To evaluate the capability of existing agents on this task, we introduce MMG2Skill-Bench, the first benchmark designed for this problem. We further propose MMG2Skill, a closed-loop framework that compiles guides into editable skills, conditions a fixed vision-language model (VLM) agent on these skills during execution, and revises the skills from trajectory-level root-cause feedback without using benchmark scores. Across GUI control, open-ended gameplay, and strategic card play with six VLM backbones, MMG2Skill consistently outperforms vanilla baseline agents in every model–domain setting, achieving macro-average gains of +12.8 to +25.3 percentage points across backbones. Ablation studies show that directly prompting agents with raw guides can degrade performance, while both structured skill construction and trajectory-driven revision are necessary for the observed improvements. On success-inferable tasks, analyzer-based early stopping further prevents late-stage performance regressions and saves 25%–53% of attempts when the success signal is properly calibrated.

Long-context capability has become increasingly critical for next-generation Large Language Models (LLMs). However, the quadratic time complexity of standard softmax attention introduces substantial computation and memory overhead as sequence length grows, limiting the scalability of LLMs in long-context scenarios. Beyond existing approaches such as linear recurrent architectures and KV-cache compression methods, we propose Kwai Summary Attention (KSA), a novel attention mechanism that reduces sequence modeling cost by compressing historical contexts into learnable summary tokens. Specifically, summary tokens are periodically inserted into the input sequence and interleaved with text tokens, partitioning the sequence into distinct chunks while enabling text tokens and summary tokens to operate under different attention scopes. This design preserves dense local expressiveness through full attention among neighboring text tokens, while modeling long-range dependencies efficiently via summary tokens. Empirical results demonstrate that hybrid-KSA consistently outperforms existing hybrid attention variants in long-context modeling. Moreover, KSA is fully orthogonal to GQA and MLA during inference, and combining KSA with these methods achieves up to 8x KV-cache compression while preserving model performance.

Leveraging the vast open-world knowledge and understanding capabilities of Large Language Models (LLMs) to develop general-purpose, semantically-aware recommender systems has emerged as a pivotal research direction in generative recommendation. However, existing methods face bottlenecks in constructing item identifiers. Text-based methods introduce LLMs’ vast output space, leading to hallucination, while methods based on Semantic IDs (SIDs) encounter a semantic gap between SIDs and LLMs’ native vocabulary, requiring costly vocabulary expansion and alignment training. To address this, this paper introduces Term IDs (TIDs), defined as a set of semantically rich and standardized textual keywords, to serve as robust item identifiers. We propose GRAM, a novel framework centered on TIDs, employs Context-aware Term Generation to convert item’s metadata into standardized TIDs and utilizes Integrative Instruction Fine-tuning to collaboratively optimize term internalization and sequential recommendation. Additionally, Elastic Identifier Grounding is designed for robust item mapping. Extensive experiments on real-world datasets demonstrate that GRAM significantly outperforms baselines across multiple scenarios, pointing a promising direction for generalizable and high-performance generative recommendation systems.

Composed Image Retrieval (CIR) enables users to search for target images using both a reference image and manipulation text, offering substantial advantages over single-modality retrieval systems. However, existing CIR methods suffer from representation space fragmentation: queries and targets comprise heterogeneous modalities and are processed by distinct encoders, forcing models to bridge misaligned representation spaces only through post-hoc alignment, which fundamentally limits retrieval performance. As evidenced by t-SNE visualization in Fig.(a), this architectural asymmetry manifests as three distinct, well-separated clusters in the feature space, directly demonstrating how heterogeneous modalities create fundamentally misaligned representation spaces from initialization. In this work, we propose CSMCIR, a unified representation framework that achieves efficient query-target alignment through three synergistic components. First, we introduce a Multi-level Chain-of-Thought (MCoT) prompting strategy that guides Multimodal Large Language Models to generate discriminative, semantically compatible captions for target images, establishing modal symmetry. Building upon this, we design a symmetric dual-tower architecture where both query and target sides utilize the identical shared-parameter Q-Former for cross-modal encoding, ensuring consistent feature representations and further reducing the alignment gap. Finally, this architectural symmetry enables an entropy-based, temporally dynamic Memory Bank strategy that provides high-quality negative samples while maintaining consistency with the evolving model state. Extensive experiments on four benchmark datasets demonstrate that our CSMCIR achieves state-of-the-art performance with superior training efficiency. Comprehensive ablation studies further validate the effectiveness of each proposed component.

The evolution of Large Language Models (LLMs) towards autonomous agents has catalyzed progress in Deep Research. While retrieval capabilities are well-benchmarked, the post-retrieval synthesis stage—where agents must digest massive amounts of context and consolidate fragmented evidence into coherent, long-form reports—remains under-evaluated due to the subjectivity of open-ended writing.To bridge this gap, we introduce DeepSynth-Eval, a benchmark designed to objectively evaluate information consolidation capabilities. We leverage high-quality survey papers as gold standards, reverse-engineer research requests, and construct Oracle Contexts from their bibliographies to isolate synthesis from retrieval noise. We propose a fine-grained evaluation protocol using General Checklists (for factual coverage) and Constraint Checklists (for structural organization), transforming subjective judgment into verifiable metrics. Experiments across 96 tasks reveal that synthesizing information from hundreds of references remains a significant challenge. Our results demonstrate that agentic “plan-then-write” workflows significantly outperform single-turn generation, effectively reducing hallucinations and improving adherence to complex structural constraints.

Long-term memory for LLM agents must answer two questions for every stored fact: whether the fact is still true and whether it should be surfaced for a given query. We refer to these two dimensions as validity and visibility, respectively. Existing memory systems often collapse these two decisions into a single scalar score. However, current and historical queries may require opposite retrieval decisions for the same fact: a superseded fact should be hidden for a current-intent query but surfaced for a historical-intent one. This creates a structural failure mode: superseded and expired facts are collapsed onto the same validity axis. To address this problem, we propose MemFSM, a training-free retrieval-side layer that separates validity from visibility through a memory-lifecycle finite-state machine. Along the validity axis, a four-state FSM tracks whether each fact has been superseded and expired. Along the visibility axis, an intent-aware router decides whether to surface or hide a fact based on its state and the query intent. By decoupling visibility from validity, MemFSM improves performance across heterogeneous memory benchmarks and diverse LLM backbones. To evaluate memory systems under fact conflicts and diverse query intents, we further introduce ValidMem, a benchmark that annotates each fact with both supersedure and expiry labels and pairs each memory store with both current-intent and historical-intent queries.

Domain specialization can improve LLM behavior, but often weakens the general capabilities inherited from the original model. Recent Multi-Teacher On-Policy Distillation (MOPD) pipelines recover model capabilities by supervising student-generated trajectories with teacher feedback, but typically assume teacher-aligned prompt coverage, requiring prompts to match the teachers’ training distributions. This assumption is difficult to satisfy when the general teacher is an open-source model whose post-training data are unknown. Instead of attempting to reconstruct this hidden distribution, we study general capability recovery with readily available proxy general prompts. We identify two failure modes of vanilla MOPD in this incomplete-coverage situation: recovery-preservation counteraction from mixing conflicting recovery and preservation gradients, and weak-signal flattening from uniformly averaging samples with unequal correction demand. We propose **Counteraction-Aware Multi-Teacher On-Policy Distillation** (**CaMOPD**), which addresses these issues with decoupled alternating training and gap-based sample selection. CaMOPD allocates dedicated updates to general recovery, periodically performs domain-preservation updates, and selects samples with larger averaged token-level teacher-student log-probability gaps to concentrate correction signals. Across role-play dialogue and medical reasoning QA scenarios, CaMOPD outperforms all baselines in general capability recovery while maintaining domain-specific behavior. Gradient coherence analyses further support the intended effect of CaMOPD in producing more coherent correction signals.

Chinese query correction (CQC) is important for search and query recommendation on content platforms, but supervised methods rely on large annotated correction pairs that are costly to maintain as query vocabularies evolve. Unsupervised correction with language models is attractive, yet in the short-query setting, unconstrained generation often over-corrects ambiguous inputs toward high-frequency phrases, causing intent drift. We propose GUIDE, a generative unsupervised framework for CQC based on a confuse-then-clarify paradigm. GUIDE encodes phonetically or visually confusable characters with shared-IDs and reconstructs the original query with an encoder–decoder architecture, which constrains correction to plausible confusion neighborhoods while learning from unlabeled query streams. A time-decayed, query-frequency-weighted objective further supports adaptation to rapidly changing query vocabularies. Experiments on QSpell 250K and KwaiSearch, a large-scale industrial dataset from Kuaishou, show that GUIDE consistently outperforms strong baselines, while online A/B testing further confirms gains in correction quality and downstream engagement.

Reinforcement learning (RL)-based enhancement of large language models (LLMs) often leads to reduced output diversity, undermining their utility in open-ended tasks like creative writing. Current methods lack explicit mechanisms for guiding diverse exploration and instead prioritize optimization efficiency and performance over diversity. This paper proposes an RL framework structured around a semi-structured long Chain-of-Thought (CoT), in which the generation process is decomposed into explicitly planned intermediate steps. We introduce a Diverse Planning Branching method that strategically introduces divergence at the planning phase based on diversity variation, alongside a group-aware diversity reward to encourage distinct trajectories. Experimental results on creative writing benchmarks demonstrate that our approach significantly improves output diversity without compromising generation quality, consistently outperforming existing baselines.

Retrieval-Augmented Generation (RAG) has emerged as a promising approach to address key limitations of Large Language Models (LLMs), such as hallucination, outdated knowledge, and lacking reliable reference. However, current RAG frameworks often struggle with identifying whether retrieved documents meaningfully contribute to answer generation. This shortcoming makes it difficult to filter out irrelevant or even misleading content, which notably impacts the final performance. In this paper, we propose Document Information Gain (DIG), a novel metric designed to quantify the contribution of retrieved documents to correct answer generation. DIG measures a document’s value by computing the difference of LLM’s generation confidence with and without the document augmented. Further, we introduce InfoGain-RAG, a framework that leverages DIG scores to train a specialized reranker, which prioritizes each retrieved document from exact distinguishing and accurate sorting perspectives. This approach can effectively filter out irrelevant documents and select the most valuable ones for better answer generation. Extensive experiments across various models and benchmarks demonstrate that InfoGain-RAG can significantly outperform existing approaches, on both single and multiple retrieval paradigm. Specifically on NaturalQA, it achieves the improvements of 17.9%, 4.5%, 12.5% in exact match accuracy against naive RAG, self-reflective RAG and modern ranking-based RAG respectively, and even an average of 15.3% increment on advanced proprietary models GPT-4o across all datasets. These results demonstrate the feasibility of InfoGain-RAG as it can offer a reliable solution for RAG in multiple applications.