Large language models (LLMs) have achieved remarkable success across a wide range of natural language processing tasks, yet their performance remains heavily biased toward high-resource languages. Tibetan, despite its cultural significance and large speaker population, is still substantially underrepresented. In this work, we present a comprehensive pipeline for advancing Tibetan language modeling through large-scale data curation and continual pre-training. We construct a 72 GB high-quality Tibetan corpus, the largest to date, and adapt Qwen2.5-7B through balanced multilingual continual pre-training with Tibetan, Chinese, and English, followed by multilingual instruction tuning. To further scale capacity efficiently, we extend the dense model to a 50B-A10B Mixture-of-Experts architecture. Due to the absence of standardized Tibetan benchmarks, we build multiple evaluation datasets via high-quality translation and human verification. Experimental results show that both dense and MoE models consistently outperform existing open-source and Tibetan-focused models of similar scale across diverse tasks. Our work advances Tibetan-centric LLM research and provides transferable insights for extending LLMs to other low-resource languages. We will release the model weights, evaluation benchmarks, and detailed data processing documentation in the follow-up.

Multimodal large language models (MLLMs) have achieved strong performance on challenging visual question answering benchmarks, yet their inference efficiency is severely constrained by the rapidly growing context. This growth stems from two primary sources: the large number of visual tokens required to encode images, and the accumulation of intermediate reasoning traces during autoregressive generation. To address these challenges, we propose LaT (**L**ook **a**nd **T**hink), the first modality-decoupled compression method that enables efficient multimodal inference. LaT structures reasoning into alternating looking and thinking steps, thereby explicitly signaling when visual grounding is required. Building on this design, LaT (1) evicts visual tokens whenever visual grounding is unnecessary, and (2) applies co-learning-guided compression after each completed step, mitigating the two sources of context growth respectively. Experimental results demonstrate that LaT reduces the average context length by up to 57%, while maintaining performance comparable to the standard MLLM baseline. The code will be publicly released.

Large language models (LLMs) have achieved impressive results in reasoning, particularly in multi-step reasoning tasks. However, when faced with more complex mathematical problems, the performance of LLMs drops significantly. To address this issue, in this paper, we propose a backward reasoning dataset, BackMATH-Data. The dataset comprises approximately 14K backward reasoning problems and 100K reasoning steps. It follows a result-oriented approach, to construct backward reasoning problems by swapping the reasoning results with specific solving conditions in the original problems.Additionally, we introduce Backward-reasoning Process-supervision Reward Model (BackPRM) and BackMATH-LLM. BackPRM supervises the quality of the generated backward reasoning problems, while BackMATH-LLM is designed for mathematical reasoning. BackMATH-LLM is fine-tuned and enhanced through reinforcement learning by supervising the quality of backward reasoning problems and by providing feedback on reasoning steps, thereby improving the mathematical reasoning capabilities of LLMs.Extensive experiments demonstrate that our model achieves an accuracy of 68.1% on the GSM8K dataset and 21.9% on the MATH dataset, exceeding the SOTA by 1.6% and 2.1% respectively.

Large language models (LLMs) have demonstrated impressive performance in reasoning. However, existing data annotation methods usually suffer from high annotation cost and the lack of effective automatic validation. To address these issues, we propose a Fine-grained Multi-Agent Debate framework (FMAD) and MMATH-Data, a dataset created by FMAD, which consists of 46K reasoning steps. By prompting multiple agents to debate, FMAD assesses the contribution of each reasoning step to the final solution, with labels based on the judge’s confidence score and the winner’s position. To facilitate reasoning in math and examine FMAD and MMATH-Data, we further propose two key components: a Multi-Agent Debate Reward Model (MRM) trained on MMATH-Data, which serves as a reward model to provide robust feedback during the optimization process, and MMATH-LLM, a model designed specifically for mathematical reasoning. MMATH-LLM is fine-tuned using reinforcement learning with supervised feedback from MRM, aiming at improving its mathematical reasoning capabilities. Extensive experiments demonstrate that our model achieves 83.4% accuracy on the GSM8K dataset and 45.1% on the MATH dataset, outperforming the state-of-the-art methods by 1.2% and 3.5%, respectively. All data and code will be available soon at GitHub.

Large language models usually suffer from multiple-file coding scenarios where strong inter-file dependencies manifest, typically demonstrated in SWE-bench. To mitigate this issue, we propose Think-Search-Patch (TSP), a retrieval-augmented reasoning framework for repository-level code repair. At the Think stage, our system breaks down a coding task and creates clear search query. Next, at the Search stage, it retrieves relevant code snippets using models like E5. At the final Patch stage, it generates standardized patches based on the key snippets. In addition the proposed framework, we enhance system reliability through a two-stage training process. At the first stage, the system undergoes supervised fine-tuning (SFT) on our TSP dataset. At the subsequent stage, we employ rejection sampling with correction to generate preference pairs for Direct Preference Optimization (DPO) training, thereby reducing errors in the intermediate phases. Experimental results demonstrate that TSP framework enhances retrieval accuracy and repair success on SWE-bench Lite, even surpassing models with a larger size in managing extensive code contexts and successfully addressing bugs spanning across multiple files. All data and code available at https://github.com/Gengar0215/TSP-framework.