1 citations · 1 across the 12 of their papers we have counts for
14 papers
On-Policy Visual Evidence Distillation
Shaohang Wei, Feifan Song, Guangyue Peng +9
Visual agents solve problems by interleaving reasoning with image operations, and on-policy distillation (OPD) provides guidance from a strong teacher on student-generated interact…
Verifier-Induced Support Reshaping in On-Policy Optimization
Shaohang Wei, Zikun Su, Feifan Song +4
We show that on-policy reinforcement learning with verifiable rewards (RLVR) can improve the current objective while making successful behaviors for later objectives too rare to sa…
Experience Augmented Policy Optimization for LLM Reasoning
Jinda Lu, Kexin Huang, Junkang Wu +7
Reinforcement Learning with Verifiable Rewards (RLVR) is a powerful paradigm for improving the reasoning capabilities of large language models (LLMs). However, existing RLVR method…
One-Way Policy Optimization for Self-Evolving LLMs
Shuo Yang, Jinda Lu, Kexin Huang +6
Reinforcement Learning with Verifiable Rewards (RLVR) has become a promising paradigm for scaling reasoning capabilities of Large Language Models (LLMs). However, the sparsity of b…
Only Say What You Know: Calibration-Aware Generation for Long-Form Factuality
Wen Luo, Guangyue Peng, Liang Wang +7
Large Reasoning Models achieve strong performance on complex tasks but remain prone to hallucinations, particularly in long-form generation where errors compound across reasoning s…
Sparse but Critical: A Token-Level Analysis of Distributional Shifts in RLVR Fine-Tuning of LLMs
Haoming Meng, Kexin Huang, Shaohang Wei +6
Reinforcement learning with verifiable rewards (RLVR) has significantly improved reasoning in large language models (LLMs), yet the token-level mechanisms underlying these improvem…
Probability-Entropy Calibration: An Elastic Indicator for Adaptive Fine-tuning
Wenhao Yu, Shaohang Wei, Jiahong Liu +5
Token-level reweighting is a simple yet effective mechanism for controlling supervised fine-tuning, but common indicators are largely one-dimensional: the ground-truth probability…
Two Pathways to Truthfulness: On the Intrinsic Encoding of LLM Hallucinations
Wen Luo, Guangyue Peng, Wei Li +8
Despite their impressive capabilities, large language models (LLMs) frequently generate hallucinations. Previous work shows that their internal states encode rich signals of truthf…
Mitigating Overthinking through Reasoning Shaping
Feifan Song, Shaohang Wei, Bofei Gao +8
Large reasoning models (LRMs) boosted by Reinforcement Learning from Verifier Reward (RLVR) have shown great power in problem solving, yet they often cause overthinking: excessive,…
Well Begun is Half Done: Low-resource Preference Alignment by Weak-to-Strong Decoding
Feifan Song, Shaohang Wei, Wen Luo +4
Large Language Models (LLMs) require alignment with human preferences to avoid generating offensive, false, or meaningless content. Recently, low-resource methods for LLM alignment…
TIME: A Multi-level Benchmark for Temporal Reasoning of LLMs in Real-World Scenarios
Shaohang Wei, Wei Li, Feifan Song +5
Temporal reasoning is pivotal for Large Language Models (LLMs) to comprehend the real world. However, existing works neglect the real-world challenges for temporal reasoning: (1) i…
Odysseus Navigates the Sirens' Song: Dynamic Focus Decoding for Factual and Diverse Open-Ended Text Generation
Wen Luo, Feifan Song, Wei Li +3
Large Language Models (LLMs) are increasingly required to generate text that is both factually accurate and diverse across various open-ended applications. However, current stochas…
CiteCheck: Towards Accurate Citation Faithfulness Detection
Ziyao Xu, Shaohang Wei, Zhuoheng Han +6
Citation faithfulness detection is critical for enhancing retrieval-augmented generation (RAG) systems, yet large-scale Chinese datasets for this task are scarce. Existing methods…
MeNTi: Bridging Medical Calculator and LLM Agent with Nested Tool Calling
Yakun Zhu, Shaohang Wei, Xu Wang +3
Integrating tools into Large Language Models (LLMs) has facilitated the widespread application. Despite this, in specialized downstream task contexts, reliance solely on tools is i…