6 papers
A Unified Per-Token Gating Family for On-Policy Distillation: FKL/RKL Mixing with Multi-Channel and Bias Coefficients
Suwan Wu, Yumeng Lin, Pengcheng Yuan +1
Per-token gating of forward/reverse KL losses has become a standard technique for on-policy knowledge distillation (OPD), but existing methods such as EOPD (Jin et al., 2026) and T…
SIRF: A Spec-Internalized Risk Foundation Model for Industrial Content Risk Control
Suwan Wu, Yumeng Lin, Pengcheng Yuan +1
For industrial content risk control, the real deployment constraint is not average accuracy but how much risk can be auto-handled under high precision and second-level latency. We…
AdaSTORM: Scaling LLM Reasoning on Dynamic Graphs via Adaptive Spatio-Temporal Multi-Agent Collaboration
Bing Hao, Ruijie Wang, Haodong Qian +5
Large Language Models (LLMs) demonstrate remarkable potential in dynamic graph reasoning, but suffer from a scaling bottleneck: current models can only handle graphs with tens of n…
SPOT: Span-level Pause-of-Thought for Efficient and Interpretable Latent Reasoning in Large Language Models
Yunlong Chu, Minglai Shao, Yuhang Liu +4
Explicit Chain-of-Thought improves the reasoning performance of large language models but often incurs high inference cost due to verbose token-level traces. While recent approache…
RouteGoT: Node-Adaptive Routing for Cost-Efficient Graph of Thoughts Reasoning
Yuhang Liu, Ruijie Wang, Yunlong Chu +4
Large Language Models (LLMs) excel at multi-step reasoning, yet increasing the structural complexity of inference does not consistently improve system-level returns. Methods such a…
Face4FairShifts: A Large Image Benchmark for Fairness and Robust Learning across Visual Domains
Yumeng Lin, Dong Li, Xintao Wu +4
Ensuring fairness and robustness in machine learning models remains a challenge, particularly under domain shifts. We present Face4FairShifts, a large-scale facial image benchmark…