In this report, we introduce DASD-4B-Thinking, a lightweight yet highly capable, fully open-source reasoning model. It achieves SOTA performance among open-source models of comparable scale across challenging benchmarks in mathematics, scientific reasoning, and code generation -- even outperforming several larger models. We begin by critically reexamining a widely adopted distillation paradigm in the community: SFT on teacher-generated responses, also known as sequence-level distillation. Although a series of recent works following this scheme have demonstrated remarkable efficiency and strong empirical performance, they are primarily grounded in the SFT perspective. Consequently, these approaches focus predominantly on designing heuristic rules for SFT data filtering, while largely overlooking the core principle of distillation itself -- enabling the student model to learn the teacher's full output distribution so as to inherit its generalization capability. Specifically, we identify three critical limitations in current practice: i) Inadequate representation of the teacher's sequence-level distribution; ii) Misalignment between the teacher's output distribution and the student's learning capacity; and iii) Exposure bias arising from teacher-forced training versus autoregressive inference. In summary, these shortcomings reflect a systemic absence of explicit teacher-student interaction throughout the distillation process, leaving the essence of distillation underexploited. To address these issues, we propose several methodological innovations that collectively form an enhanced sequence-level distillation training pipeline. Remarkably, DASD-4B-Thinking obtains competitive results using only 448K training samples -- an order of magnitude fewer than those employed by most existing open-source efforts. To support community research, we publicly release our models and the training dataset.
MedCLM: Learning to Localize and Reason via a CoT-Curriculum in Medical Vision-Language Models
A.I.MATICS Inc(A.I.MATICS公司);Boston University(波士顿大学);University of Southern California(南加州大学);Heuron(Heuron公司);MODULABS, Open Neural Networks Research Lab(MODULABS,开放神经网络研究实验室)
DSBD: Dual-Aligned Structural Basis Distillation for Graph Domain Adaptation
Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学);Zhengzhou University(郑州大学);The Hebei University of Technology(河北工业大学);California State Polytechnic University(加州州立理工大学);The City University of Hong Kong(香港城市大学)
Multi-hop Deep Joint Source-Channel Coding with Deep Hash Distillation for Semantically Aligned Image Recovery
Department of Electrical and Electronic Engineering, Imperial College London, UK(电气与电子工程系,帝国理工学院伦敦分校,英国);Lehrstuhl für Nachrichtentechnik, Technische Universität Dortmund, Germany(信息论系,德意志理工大学多特蒙德分校,德国)
Frequency-Aligned Knowledge Distillation for Lightweight Spatiotemporal Forecasting
Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所);The University of Hong Kong(香港大学);The City University of New York(纽约城市大学);Tsinghua University(清华大学)
Beyond Timestamps: Decision-Aligned On-Policy Distillation for Long-Horizon Agents
Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing(北京未来区块链与隐私计算高精尖创新中心);Tsinghua University(清华大学);Beihang University(北京航空航天大学)
EvoCoT: Overcoming the Exploration Bottleneck in Reinforcement Learning
School of Computer Science, Peking University(北京大学计算机科学学院);College of AI, Tsinghua University(清华大学人工智能学院);Imperial College London(伦敦帝国理工学院);Institute of Software, Chinese Academy of Sciences(中国科学院软件研究所)
Do Latent-CoT Models Think Step-by-Step? A Mechanistic Study on Sequential Reasoning Tasks
Institute for Computational and Mathematical Engineering, Stanford University, Stanford, USA(计算与数学工程研究所,斯坦福大学);MOE Key Lab of Computational Linguistics, Peking University, Beijing, China(教育部计算语言学重点实验室,北京大学);Beijing Academy of Artificial Intelligence, Beijing, China(北京人工智能研究院)