4 papers
From Search to Synthesis: Training LLMs as Zero-Shot Workflow Generators
Gan Luo, Zihan Qin, Bin Dong +1
Large language models (LLMs) excel across a wide range of tasks, yet their instance-specific solutions often lack the structural consistency needed for reliable deployment. Workflo…
Exploration vs Exploitation: Rethinking RLVR through Clipping, Entropy, and Spurious Reward
Peter Chen, Xiaopeng Li, Ziniu Li +3
This paper examines the exploration-exploitation trade-off in reinforcement learning with verifiable rewards (RLVR), a framework for improving the reasoning of Large Language Model…
ComPO: Preference Alignment via Comparison Oracles
Peter Chen, Xi Chen, Wotao Yin +1
Direct alignment methods are increasingly used for aligning large language models (LLMs) with human preferences. However, these methods suffer from the issues of verbosity and like…
Block Acceleration Without Momentum: On Optimal Stepsizes of Block Gradient Descent for Least-Squares
Liangzu Peng, Wotao Yin
Block coordinate descent is a powerful algorithmic template suitable for big data optimization. This template admits a lot of variants including block gradient descent (BGD), which…