12 papers
Unifying Group-Relative and Self-Distillation Policy Optimization via Sample Routing
Gengsheng Li, Tianyu Yang, Junfeng Fang +6
Reinforcement learning with verifiable rewards (RLVR) has become a standard paradigm for post-training large language models. While Group Relative Policy Optimization (GRPO) is wid…
PRISM: Probability Reallocation with In-Span Masking for Knowledge-Sensitive Alignment
Chenning Xu, Mao Zheng, Mingyang Song
Supervised fine-tuning (SFT) with token-level hard labels can amplify overconfident imitation of factually unsupported targets, causing hallucinations that propagate in multi-sente…
Model Merging in the Era of Large Language Models: Methods, Applications, and Future Directions
Mingyang Song, Mao Zheng
Model merging combines the parameters of multiple neural networks into a single model without additional training. As fine-tuned large language models (LLMs) proliferate, merging o…
Beyond the Illusion of Consensus: From Surface Heuristics to Knowledge-Grounded Evaluation in LLM-as-a-Judge
Mingyang Song, Mao Zheng, Chenning Xu
The paradigm of LLM-as-a-judge relies on a critical assumption, namely that high inter-evaluator agreement indicates reliable and objective evaluation. We present two complementary…
CodeDelegator: Mitigating Context Pollution via Role Separation in Code-as-Action Agents
Tianxiang Fei, Cheng Chen, Yue Pan +2
Recent advances in large language models (LLMs) allow agents to represent actions as executable code, offering greater expressivity than traditional tool-calling. However, real-wor…
PodBench: A Comprehensive Benchmark for Instruction-Aware Audio-Oriented Podcast Script Generation
Chenning Xu, Mao Zheng, Mingyu Zheng +1
Podcast script generation requires LLMs to synthesize structured, context-grounded dialogue from diverse inputs, yet systematic evaluation resources for this task remain limited. T…