Ruobing Xie

Efficient Reasoning Exploration via State-Conditioned Latent Steering with Progress Guidance

Hengyuan Zhang, Chenming Shang, Zunhai Su +10

Best-of- $N$ is a widely used inference strategy for complex reasoning, whose effectiveness depends on whether sampled candidates can cover diverse and high-quality reasoning paths.…

Hybrid Gated Attention

Zekun Zhou, Ruobing Xie, Lanrui Wang +1

Gated attention is an effective approach to mitigate attention sinks and enhance the representational capacity of attention. To further extend its effectiveness-efficiency Pareto f…

AngelSpec: Towards Real-World High Performance Inference with Speculative Decoding

Hong Liu, Rui Cen, Junhan Shi +10

Speculative decoding accelerates large language model inference without changing the target distribution, but no single drafting structure performs best across real-world workloads…

Negative Advantages Is a Double-Edged Sword: Calibrating advantages in GRPO for Search Agents

Jiayi Wu, Ruobing Xie, Zeqian Huang +6

Search agents achieve strong question-answering performance through multi-turn interactions with search engines, with Group Relative Policy Optimization (GRPO) being a widely used…

Self-Distillation for Multi-Token Prediction

Guoliang Zhao, Ruobing Xie, An Wang +3

As Large Language Models (LLMs) scale up, inference efficiency becomes a critical bottleneck. Multi-Token Prediction (MTP) could accelerate LLM inference by predicting multiple fut…

LaSeR: Reinforcement Learning with Last-Token Self-Rewarding

Wenkai Yang, Weijie Liu, Ruobing Xie +4

Reinforcement Learning with Verifiable Rewards (RLVR) has recently emerged as a core paradigm for enhancing the reasoning capabilities of Large Language Models (LLMs). To address t…

Flexible Realignment of Language Models

Wenhong Zhu, Ruobing Xie, Weinan Zhang +1

Realignment becomes necessary when a language model (LM) fails to meet expected performance. We propose a flexible realignment framework that supports quantitative control of align…

The Climb Carves Wisdom Deeper Than the Summit: On the Noisy Rewards in Learning to Reason

Ang Lv, Ruobing Xie, Xingwu Sun +2

Recent studies on post-training large language models (LLMs) for reasoning through reinforcement learning (RL) typically focus on tasks that can be accurately verified and rewarded…

Hunyuan-TurboS: Advancing Large Language Models through Mamba-Transformer Synergy and Adaptive Chain-of-Thought

Tencent Hunyuan Team, Ao Liu, Botong Zhou +248

As Large Language Models (LLMs) rapidly advance, we introduce Hunyuan-TurboS, a novel large hybrid Transformer-Mamba Mixture of Experts (MoE) model. It synergistically combines Mam…

Mitigating Hallucination in Multimodal Large Language Model via Hallucination-targeted Direct Preference Optimization

Yuhan Fu, Ruobing Xie, Xingwu Sun +2

Multimodal Large Language Models (MLLMs) are known to hallucinate, which limits their practical applications. Recent works have attempted to apply Direct Preference Optimization (D…

cs.CL2024★ 1 cited

More Expressive Attention with Negative Weights

Ang Lv, Ruobing Xie, Shuaipeng Li +5

We propose a novel attention mechanism, named Cog Attention, that enables attention weights to be negative for enhanced expressiveness, which stems from two key factors: (1) Cog At…

Exploring Forgetting in Large Language Model Pre-Training

Chonghua Liao, Ruobing Xie, Xingwu Sun +2

Catastrophic forgetting remains a formidable obstacle to building an omniscient model in large language models (LLMs). Despite the pioneering research on task-level forgetting in L…

Lossless KV Cache Compression to 2%

Zhen Yang, J. N. Han, Kan Wu +4

Large language models have revolutionized data processing in numerous domains, with their ability to handle extended context reasoning receiving notable recognition. To speed up in…