An Empirical Study and Theoretical Explanation on Task-Level Model-Merging Collapse

Model merging unifies independently fine-tuned LLMs from the same base, enabling reuse and integration of parallel development efforts without retraining. However, in practice we observe that merging does not always succeed: certain combinations of task-specialist models suffer from catastrophic performance degradation after merging. We refer to this failure mode as merging collapse. Intuitively, collapse arises when the learned representations or parameter adjustments for different tasks are fundamentally incompatible, so that merging forces destructive interference rather than synergy. In this paper, we identify and characterize the phenomenon of task-level merging collapse, where certain task combinations consistently trigger huge performance degradation across all merging methods. Through extensive experiments and statistical analysis, we demonstrate that representational incompatibility between tasks is strongly correlated with merging collapse, while parameter-space conflict metrics show minimal correlation, challenging conventional wisdom in model merging literature. We provide a theoretical explanation on this phenomenon through rate-distortion theory with a dimension-dependent bound, establishing fundamental limits on task mergeability regardless of methodology.

LiDAR Prompted Spatio-Temporal Multi-View Stereo for Autonomous Driving

Accurate metric depth is critical for autonomous driving perception and simulation, yet current approaches struggle to achieve high metric accuracy, multi-view and temporal consistency, and cross-domain generalization. To address these challenges, we present DriveMVS, a novel multi-view stereo framework that reconciles these competing objectives through two key insights: (1) Sparse but metrically accurate LiDAR observations can serve as geometric prompts to anchor depth estimation in absolute scale, and (2) deep fusion of diverse cues is essential for resolving ambiguities and enhancing robustness, while a spatio-temporal decoder ensures consistency across frames. Built upon these principles, DriveMVS embeds the LiDAR prompt in two ways: as a hard geometric prior that anchors the cost volume, and as soft feature-wise guidance fused by a triple-cue combiner. Regarding temporal consistency, DriveMVS employs a spatio-temporal decoder that jointly leverages geometric cues from the MVS cost volume and temporal context from neighboring frames. Experiments show that DriveMVS achieves state-of-the-art performance on multiple benchmarks, excelling in metric accuracy, temporal stability, and zero-shot cross-domain transfer, demonstrating its practical value for scalable, reliable autonomous driving systems.

GUI-GENESIS: Automated Synthesis of Efficient Environments with Verifiable Rewards for GUI Agent Post-Training

Post-training GUI agents in interactive environments is critical for developing generalization and long-horizon planning capabilities. However, training on real-world applications is hindered by high latency, poor reproducibility, and unverifiable rewards relying on noisy visual proxies. To address the limitations, we present GUI-GENESIS, the first framework to automatically synthesize efficient GUI training environments with verifiable rewards. GUI-GENESIS reconstructs real-world applications into lightweight web environments using multimodal code models and equips them with code-native rewards, executable assertions that provide deterministic reward signals and eliminate visual estimation noise. Extensive experiments show that GUI-GENESIS reduces environment latency by 10 times and costs by over $28,000 per epoch compared to training on real applications. Notably, agents trained with GUI-GENESIS outperform the base model by 14.54% and even real-world RL baselines by 3.27% on held-out real-world tasks. Finally, we observe that models can synthesize environments they cannot yet solve, highlighting a pathway for self-improving agents.

LongStream: Long-Sequence Streaming Autoregressive Visual Geometry

Long-sequence streaming 3D reconstruction remains a significant open challenge. Existing autoregressive models often fail when processing long sequences. They typically anchor poses to the first frame, which leads to attention decay, scale drift, and extrapolation errors. We introduce LongStream, a novel gauge-decoupled streaming visual geometry model for metric-scale scene reconstruction across thousands of frames. Our approach is threefold. First, we discard the first-frame anchor and predict keyframe-relative poses. This reformulates long-range extrapolation into a constant-difficulty local task. Second, we introduce orthogonal scale learning. This method fully disentangles geometry from scale estimation to suppress drift. Finally, we solve Transformer cache issues such as attention-sink reliance and long-term KV-cache contamination. We propose cache-consistent training combined with periodic cache refresh. This approach suppresses attention degradation over ultra-long sequences and reduces the gap between training and inference. Experiments show LongStream achieves state-of-the-art performance. It delivers stable, metric-scale reconstruction over kilometer-scale sequences at 18 FPS. Project Page: https://3dagentworld.github.io/longstream/

TFEC: Multivariate Time-Series Clustering via Temporal-Frequency Enhanced Contrastive Learning

Multivariate Time-Series (MTS) clustering is crucial for signal processing and data analysis. Although deep learning approaches, particularly those leveraging Contrastive Learning (CL), are prominent for MTS representation, existing CL-based models face two key limitations: 1) neglecting clustering information during positive/negative sample pair construction, and 2) introducing unreasonable inductive biases, e.g., destroying time dependence and periodicity through augmentation strategies, compromising representation quality. This paper, therefore, proposes a Temporal-Frequency Enhanced Contrastive (TFEC) learning framework. To preserve temporal structure while generating low-distortion representations, a temporal-frequency Co-EnHancement (CoEH) mechanism is introduced. Accordingly, a synergistic dual-path representation and cluster distribution learning framework is designed to jointly optimize cluster structure and representation fidelity. Experiments on six real-world benchmark datasets demonstrate TFEC's superiority, achieving 4.48% average NMI gains over SOTA methods, with ablation studies validating the design. The code of the paper is available at: https://github.com/yueliangy/TFEC.

Lil: Less is Less When Applying Post-Training Sparse-Attention Algorithms in Long-Decode Stage

Large language models (LLMs) demonstrate strong capabilities across a wide range of complex tasks and are increasingly deployed at scale, placing significant demands on inference efficiency. Prior work typically decomposes inference into prefill and decode stages, with the decode stage dominating total latency. To reduce time and memory complexity in the decode stage, a line of work introduces sparse-attention algorithms. In this paper, we show, both empirically and theoretically, that sparse attention can paradoxically increase end-to-end complexity: information loss often induces significantly longer sequences, a phenomenon we term ``Less is Less'' (Lil). To mitigate the Lil problem, we propose an early-stopping algorithm that detects the threshold where information loss exceeds information gain during sparse decoding. Our early-stopping algorithm reduces token consumption by up to 90% with a marginal accuracy degradation of less than 2% across reasoning-intensive benchmarks.

WebCoderBench: Benchmarking Web Application Generation with Comprehensive and Interpretable Evaluation Metrics

Web applications (web apps) have become a key arena for large language models (LLMs) to demonstrate their code generation capabilities and commercial potential. However, building a benchmark for LLM-generated web apps remains challenging due to the need for real-world user requirements, generalizable evaluation metrics without relying on ground-truth implementations or test cases, and interpretable evaluation results. To address these challenges, we introduce WebCoderBench, the first real-world-collected, generalizable, and interpretable benchmark for web app generation. WebCoderBench comprises 1,572 real user requirements, covering diverse modalities and expression styles that reflect realistic user intentions. WebCoderBench provides 24 fine-grained evaluation metrics across 9 perspectives, combining rule-based and LLM-as-a-judge paradigm for fully automated, objective, and general evaluation. Moreover, WebCoderBench adopts human-preference-aligned weights over metrics to yield interpretable overall scores. Experiments across 12 representative LLMs and 2 LLM-based agents show that there exists no dominant model across all evaluation metrics, offering an opportunity for LLM developers to optimize their models in a targeted manner for a more powerful version.

From User Interface to Agent Interface: Efficiency Optimization of UI Representations for LLM Agents

While Large Language Model (LLM) agents show great potential for automated UI navigation such as automated UI testing and AI assistants, their efficiency has been largely overlooked. Our motivating study reveals that inefficient UI representation creates a critical performance bottleneck. However, UI representation optimization, formulated as the task of automatically generating programs that transform UI representations, faces two unique challenges. First, the lack of Boolean oracles, which traditional program synthesis uses to decisively validate semantic correctness, poses a fundamental challenge to co-optimization of token efficiency and completeness. Second, the need to process large, complex UI trees as input while generating long, compositional transformation programs, making the search space vast and error-prone. Toward addressing the preceding limitations, we present UIFormer, the first automated optimization framework that synthesizes UI transformation programs by conducting constraint-based optimization with structured decomposition of the complex synthesis task. First, UIFormer restricts the program space using a domain-specific language (DSL) that captures UI-specific operations. Second, UIFormer conducts LLM-based iterative refinement with correctness and efficiency rewards, providing guidance for achieving the efficiency-completeness co-optimization. UIFormer operates as a lightweight plugin that applies transformation programs for seamless integration with existing LLM agents, requiring minimal modifications to their core logic. Evaluations across three UI navigation benchmarks spanning Android and Web platforms with five LLMs demonstrate that UIFormer achieves 48.7% to 55.8% token reduction with minimal runtime overhead while maintaining or improving agent performance. Real-world industry deployment at WeChat further validates the practical impact of UIFormer.

Text2SQL-Flow: A Robust SQL-Aware Data Augmentation Framework for Text-to-SQL

The data-centric paradigm has become pivotal in AI, especially for Text-to-SQL, where performance is limited by scarce, simplistic, and low-diversity datasets. To address this, we propose Text2SQL-Flow, a SQL-aware data augmentation framework that generates large-scale, semantically valid, and structurally diverse Text-to-SQL pairs from minimal seed data. It operates across six augmentation dimensions and integrates an end-to-end pipeline featuring SQL execution verification, natural language question generation, chain-of-thought reasoning traces, and data classification. A modular Database Manager ensures cross-database compatibility and scalability. Using this framework, we build SQLFlow, a high-quality dataset of 89,544 annotated examples. We evaluate SQLFlow in two settings: (1) For open-source LLMs, fine-tuning on SQLFlow consistently improves performance across benchmarks under the same data budget. (2) For closed-source LLMs, we introduce a masked alignment retrieval method that treats SQLFlow as both knowledge base and training data for the retriever. This enables structure-aware example matching by modeling fine-grained alignments between questions and SQL queries. Experiments show our retrieval strategy outperforms existing methods, underscoring the value of SQLFlow's high-fidelity data and our novel technique. Our work establishes a scalable, data-centric foundation for advancing Text-to-SQL systems and highlights the critical role of high-quality structured data in modern AI.

AppForge: From Assistant to Independent Developer - Are GPTs Ready for Software Development?

Large language models (LLMs) have demonstrated remarkable capability in function-level code generation tasks. Unlike isolated functions, real-world applications demand reasoning over the entire software system: developers must orchestrate how different components interact, maintain consistency across states over time, and ensure the application behaves correctly within the lifecycle and framework constraints. Yet, no existing benchmark adequately evaluates whether LLMs can bridge this gap and construct entire software systems from scratch. To address this gap, we propose APPFORGE, a benchmark consisting of 101 software development problems drawn from real-world Android apps. Given a natural language specification detailing the app functionality, a language model is tasked with implementing the functionality into an Android app from scratch. Developing an Android app from scratch requires understanding and coordinating app states, lifecycle management, and asynchronous operations, calling for LLMs to generate context-aware, robust, and maintainable code. To construct APPFORGE, we design a multi-agent system to automatically summarize the main functionalities from app documents and navigate the app to synthesize test cases validating the functional correctness of app implementation. Following rigorous manual verification by Android development experts, APPFORGE incorporates the test cases within an automated evaluation framework that enables reproducible assessment without human intervention, making it easily adoptable for future research. Our evaluation on 12 flagship LLMs show that all evaluated models achieve low effectiveness, with the best-performing model (GPT-5) developing only 18.8% functionally correct applications, highlighting fundamental limitations in current models' ability to handle complex, multi-component software engineering challenges.