22 papers
WDL-OPD: Weak-Driven On-Policy Distillation via Mixture-Constrained Co-Training
Zehao Chen, Gongxun Li, Tianxiang Ai +9
On-policy distillation (OPD) aligns a student with a teacher on trajectories sampled from the student itself, reducing the train-test state mismatch of offline distillation. The sa…
CodeChemist: Test-Time Scaling for Low-Resource Code Generation via Functional Knowledge Transfer
Kaixin Wang, Tianlin Li, Xiaoyu Zhang +6
Code Large Language Models (CodeLLMs) have been widely adopted for Natural Language to Programming Language code generation, powering applications with large user bases. Their perf…
StainFlow: Entity-Stain Tracking and Evidence Linking for Process Rewards in GUI Agents
Haojie Hao, Longkun Hao, Yihang Lou +8
Reinforcement Learning (RL) has become a promising approach for improving GUI Agents in long-horizon, stochastic digital environments, but trajectory-level success feedback is too…
Weak-Driven Learning: How Weak Agents make Strong Agents Stronger
Zehao Chen, Gongxun Li, Tianxiang Ai +9
As post-training optimization becomes central to improving large language models, we observe a persistent saturation bottleneck: once models grow highly confident, further training…
MIRA: Mid-training Rubric Anchoring for Source-Aware Data Selection
Haowen Wang, Yaxin Du, Jian Yang +9
Mid-training has become an important stage in modern LLM development, using large-scale curated mixtures to strengthen capabilities before final post-training. Its data selection p…
SGMD: Score Gradient Matching Distillation for Few-Step Video Diffusion Distillation
Zhuguanyu Wu, Ruihao Gong, Yang Yong +5
Distribution Matching Distillation (DMD) is a widely used paradigm for accelerating inference in few-step video diffusion models. However, DMD-style video distillation faces two co…