Recent research reveals that a minority of high-entropy tokens significantly influence the reasoning quality of large language models (LLMs). Inspired by this, we propose Prototype Entropy Alignment (PEA), a reinforcement learning framework that models effective reasoning not as a single path but as a collection of learnable "entropy signatures." PEA identifies these signatures by clustering expert trajectories' uncertainty patterns into a diverse and continuously updated set of prototypes. The model is then rewarded for aligning its own reasoning process with these evolving targets, creating a self-improvement loop. Instead of replacing traditional outcome-based rewards, PEA provides a complementary, process-oriented signal. Our experiments show that this synergy is crucial: PEA substantially boosts performance on creative and general reasoning tasks and, when combined with outcome rewards, achieves SOTA results on structured tasks such as mathematics. By rewarding alignment with diverse and evolving reasoning structures, PEA offers a robust, verifier-free pathway to enhance reasoning's adaptability.

Guanran Luo, Wentao Qiu, Wanru Zhao, Wenhan Lv, Zhongquan Jian, Meihong Wang, Qingqiang Wu. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Recent advances in code generation have illuminated the potential of employing large language models (LLMs) for general-purpose programming languages such as Python and C++, opening new opportunities for automating software development and enhancing programmer productivity. The potential of LLMs in software programming has sparked significant interest in exploring automated hardware generation and automation. Although preliminary endeavors have been made to adopt LLMs in generating hardware description languages (HDLs) such as Verilog and SystemVerilog, several challenges persist in this direction. First, the volume of available HDL training data is substantially smaller compared to that for software programming languages. Second, the pre-trained LLMs, mainly tailored for software code, tend to produce HDL designs that are more error-prone. Third, the generation of HDL requires a significantly higher number of tokens compared to software programming, leading to inefficiencies in cost and energy consumption. To tackle these challenges, this paper explores leveraging LLMs to generate High-Level Synthesis (HLS)-based hardware design. Although code generation for domain-specific programming languages is not new in the literature, we aim to provide experimental results, insights, benchmarks, and evaluation infrastructure to investigate the suitability of HLS over low-level HDLs for LLM-assisted hardware design generation. To achieve this, we first finetune pretrained models for HLS-based hardware generation, using a collected dataset with text prompts and corresponding reference HLS designs. An LLM-assisted framework is then proposed to automate end-to-end hardware code generation, which also investigates the impact of chain-of-thought and feedback loops promoting techniques on HLS- design generation. Comprehensive experiments demonstrate the effectiveness of our methods.

Curriculum learning is a class of training strategies that organizes the data being exposed to a model by difficulty, gradually from simpler to more complex examples. This research explores a reverse curriculum generation approach that recursively decomposes complex datasets into simpler, more learnable components. We propose a teacher-student framework where the teacher is equipped with the ability to reason step-by-step, which is used to recursively generate easier versions of examples, enabling the student model to progressively master difficult tasks. We propose a novel scoring system to measure data difficulty based on its structural complexity and conceptual depth, allowing curriculum construction over decomposed data. Experiments on math datasets (MATH and AIME) and code generation datasets demonstrate that models trained with curricula generated by our approach exhibit superior performance compared to standard training on original datasets.

Pretrained large language models (LLMs) have emerged as a cornerstone in modern natural language processing, with their utility expanding to various applications and languages. However, the fine-tuning of multilingual LLMs, particularly for low-resource languages, is fraught with challenges steming from data-sharing restrictions (the physical border) and from the inherent linguistic differences (the linguistic border). These barriers hinder users of various languages, especially those in low-resource regions, from fully benefiting from the advantages of LLMs. To overcome these challenges, we propose the Federated Prompt Tuning Paradigm for Multilingual Scenarios, which leverages parameter-efficient fine-tuning in a manner that preserves user privacy. We have designed a comprehensive set of experiments and introduced the concept of "language distance" to highlight the several strengths of this paradigm. Even under computational constraints, our method not only bolsters data efficiency but also facilitates mutual enhancements across languages, particularly benefiting low-resource ones. Compared to traditional local crosslingual transfer tuning methods, our approach achieves a 6.9\% higher accuracy, reduces the training parameters by over 99\%, and demonstrates stronger cross-lingual generalization. Such findings underscore the potential of our approach to promote social equality, ensure user privacy, and champion linguistic diversity.

Large language models (LLMs) have shown increasing capability in problem-solving and decision-making, largely based on the step-by-step chain-of-thought reasoning processes. However, evaluating these reasoning abilities has become increasingly challenging. Existing outcome-based benchmarks are beginning to saturate, becoming less effective in tracking meaningful progress. To address this, we present a process-based benchmark MR-Ben that demands a meta-reasoning skill, where LMs are asked to locate and analyse potential errors in automatically generated reasoning steps. Our meta-reasoning paradigm is especially suited for system-2 slow thinking, mirroring the human cognitive process of carefully examining assumptions, conditions, calculations, and logic to identify mistakes. MR-Ben comprises 5,975 questions curated by human experts across a wide range of subjects, including physics, chemistry, logic, coding, and more. Through our designed metrics for assessing meta-reasoning on this benchmark, we identify interesting limitations and weaknesses of current LLMs (open-source and closed-source models). For example, with models like the o1 series from OpenAI demonstrating strong performance by effectively scrutinizing the solution space, many other state-of-the-art models fall significantly behind on MR-Ben, exposing potential shortcomings in their training strategies and inference methodologies.

Recent research has highlighted the importance of data quality in scaling large language models (LLMs). However, automated data quality control faces unique challenges in collaborative settings where sharing is not allowed directly between data silos. To tackle this issue, this paper proposes a novel data quality control technique based on the notion of data influence on the training dynamics of LLMs, that high quality data are more likely to have similar training dynamics to the anchor dataset. We then leverage the influence of the training dynamics to select high-quality data from different private domains, with centralized model updates on the server side in a collaborative training fashion by either model merging or federated learning. As for the data quality indicator, we compute the per-sample gradients with respect to the private data and the anchor dataset, and use the trace of the accumulated inner products as a measurement of data quality. In addition, we develop a quality control evaluation tailored for collaborative settings with heterogeneous medical domain data. Experiments show that training on the high-quality data selected by our method can often outperform other data selection methods for collaborative fine-tuning of LLMs, across diverse private domain datasets, in medical, multilingual and financial settings.

The recent advances in 3D Convolutional Neural Networks (3D CNNs) have shown promising performance for untrimmed video action detection, employing the popular detection framework that heavily relies on the temporal action proposal generations as the input of the action detector and localization regressor. In practice the proposals usually contain strong intra and inter relations among them, mainly stemming from the temporal and spatial variations in the video actions. However, most of existing 3D CNNs ignore the relations and thus suffer from the redundant proposals degenerating the detection performance and efficiency. To address this problem, we propose graph attention based proposal 3D ConvNets (AGCN-P-3DCNNs) for video action detection. Specifically, our proposed graph attention is composed of intra attention based GCN and inter attention based GCN. We use intra attention to learn the intra long-range dependencies inside each action proposal and update node matrix of Intra Attention based GCN, and use inter attention to learn the inter dependencies between different action proposals as adjacency matrix of Inter Attention based GCN. Afterwards, we fuse intra and inter attention to model intra long-range dependencies and inter dependencies simultaneously. Another contribution is that we propose a simple and effective framewise classifier, which enhances the feature presentation capabilities of backbone model. Experiments on two proposal 3D ConvNets based models (P-C3D and P-ResNet) and two popular action detection benchmarks (THUMOS 2014, ActivityNet v1.3) demonstrate the state-of-the-art performance achieved by our method. Particularly, P-C3D embedded with our module achieves average mAP 3.7% improvement on THUMOS 2014 dataset compared to original model.