Autoregressive vision-language models achieve strong multimodal understanding, but their sequential decoding becomes increasingly inefficient as multimodal responses grow longer. Adapting AR-VLMs to diffusion-style block decoding is promising, yet fixed refinement blocks struggle with sharply varying multimodal uncertainty in OCR, chart, document, and reasoning-heavy outputs. Binary mask-or-token states further limit local refinement, as masking discards partial evidence while premature commitment can corrupt the causal prefix. We present FLUID-V, a lightweight AR-to-diffusion adaptation framework for reliable local multimodal refinement. FLUID-V selects refinement spans with elastic horizons and represents unresolved positions as evolving soft semantic mixtures in embedding space. Only confident predictions are committed to the persistent KV cache, preserving causal multimodal conditioning and autoregressive compatibility. Experiments with OpenPangu-VL-7B and 100K self-distilled samples show competitive performance across diverse multimodal benchmarks, with ablations confirming the importance of adaptive horizons and stateful intermediate representations. Our code is available at https://github.com/Oli-lab-nun/FLUID-V.

Diffusion models promise efficient parallel text generation but rely on bidirectional attention, creating a structural mismatch with pre-trained Autoregressive (AR) models. This incompatibility precludes reusing robust AR priors, necessitating prohibitive pre-training from scratch. To bridge this gap, we propose FLUID, a framework that efficiently adapts AR backbones to the diffusion paradigm. By enforcing Strictly Causal Alignment, FLUID enables seamless initialization from standard GPT-style checkpoints, circumventing the need for massive pre-training. Furthermore, we introduce Elastic Horizons, an entropy-driven mechanism that dynamically modulates denoising strides based on local information density rather than fixed schedules. Experiments demonstrate that FLUID achieves state-of-the-art performance while reducing training costs by orders of magnitude, effectively reconciling established AR foundations with efficient parallel generation. Our code is available at https://huggingface.co/MYTH-Lab/FLUID.

Post-training quantization facilitates efficient LLM deployment, yet it disproportionately erodes reasoning and mathematical performance. To recover fidelity without costly fine-tuning, we identify a critical vulnerability in existing output-space methods: the tendency to overfit the limited calibration manifold, which necessitates structural regularization for unseen data. We propose **ROCOB** (**R**esilient **O**utput **C**ompensation with **O**ptimal **B**udgeting), a training-free framework based on *Dual-Manifold Alignment*. ROCOB couples a Principal Activation Component (PAC), which minimizes error on the calibration manifold, with an Orthogonal Geometric Component (OGC), which recovers residual weight-space structure without fitting a second calibration objective. To compare methods fairly under the same added-parameter budget, ROCOB uses Global Spectral Energy Maximization (GSEM) to allocate ranks according to recovered spectral energy and module cost. Across Qwen3 and Llama-3 models, quantizers, and reasoning benchmarks, ROCOB improves average post-quantization recovery in most evaluated model–quantizer settings while preserving a training-free construction pipeline.

Large Language Models (LLMs) and Retrieval Augmented Generation (RAG) methods have demonstrated significant potential on tasks across multiple domains. However, ellipses and coreferences, as common phenomena in dialogue scenes, pose challenges to LLMs’ understanding and RAG’s retrieval accuracy. The previous works ignore the negative impact of this fuzzy data on RAG system.We explore the capabilities of LLMs and RAG systems in dialogue scenarios and use Incomplete Utterance Rewriting (IUR) to complete the key information in dialogue to enhance retrieval.Besides, we propose a lightweight IUR model for query rewriting. It is an end-to-end framework for node linking and iterative inference, incorporating two newly proposed probing semantic features derived from generative pre-training. This framework treats IUR as a series of link decisions on the input sequence and the incrementally constructed rewriting outputs.To test the performance of RAG system in the model multi-round dialogue scenario, we construct an RAG dialogue dataset on English and Chinese, Dialogue-RAG-MULTI-v1.0.Experiment results show that utterance rewriting can effectively improve the retrieval and generation ability of RAG system in dialogue scenes. Experiments on IUR tasks demonstrate the excellent performance of our lightweight IUR method.