4 papers
To Reason or to Fabricate: Reasoning Without Shortcuts via Hint-Anchored Pairwise Aggregation
Jiuheng Lin, Chen Zhang, Yansong Feng
While reinforcement learning (RL) significantly enhances LLM reasoning, its efficacy is severely undermined by Pre-RL data overlap, where RL datasets overlap with pretraining or SF…
EVPO: Explained Variance Policy Optimization for Adaptive Critic Utilization in LLM Post-Training
Chengjun Pan, Shichun Liu, Jiahang Lin +10
Reinforcement learning (RL) for LLM post-training faces a fundamental design choice: whether to use a learned critic as a baseline for policy optimization. Classical theory favors…
Efficient Low-Resource Language Adaptation via Multi-Source Dynamic Logit Fusion
Chen Zhang, Jiuheng Lin, Zhiyuan Liao +1
Adapting large language models (LLMs) to low-resource languages (LRLs) is constrained by the scarcity of task data and computational resources. Although Proxy Tuning offers a logit…
How Much Do Large Language Models Know about Human Motion? A Case Study in 3D Avatar Control
Kunhang Li, Jason Naradowsky, Yansong Feng +1
We explore the human motion knowledge of Large Language Models (LLMs) through 3D avatar control. Given a motion instruction, we prompt LLMs to first generate a high-level movement…