57 citations · 715 across the 196 of their papers we have counts for
160 papers · 1 filter
Meet Students Where They Are: Student-Prefix Teacher Adaptation for On-Policy Distillation
Xue Zhang, Songming Zhang, Yunlong Liang +3
On-policy distillation (OPD) mitigates the student's exposure bias by optimizing on student-generated trajectories. However, student-generated prefixes may contain reasoning errors…
From Candidate Tokens to Final Outcomes: Decoding Relative Long-Horizon Candidate Value from Reasoning Model Representations
Yixin Yang, Yue Wang, Fandong Meng +1
Large reasoning models generate reasoning trajectories token by token. At a given prefix, multiple next-token candidates may be plausible, yet shape subsequent reasoning and differ…
Rewrite What Matters: Adaptive Multilingual Query Rewriting for Reasoning via Agentic Reinforcement Learning
Rui Qi, Yufeng Chen, Yunlong Liang +6
In multilingual scenarios, queries with equivalent semantics but in different languages could guide the model into different reasoning trajectories, leading to performance disparit…
Not Just Reason, Not Just Scan: Reinforcement Learning for Proactive Scientific Error Verification over Academic Paper
Rongjin Li, Yuanxin Liu, Hao Zhou +3
Multimodal large language models (MLLMs) are increasingly capable scientific assistants, yet they remain far from fully autonomous research. This transition requires models to acti…
Self-Improving Large Language Models via Progressive Experience Evolution
Shijie Ren, Xiting Wang, Meng Li +8
Large language models (LLMs) capable of self-improvement require not only effective policy optimization, but also a principled mechanism for transforming transient interaction expe…
Knowing When to Quit: Diagnosing and Training LLMs to Abort Futile Reasoning
Xinyan Guan, Jiali Zeng, Chunlei Xin +5
Large language models generate computationally expensive yet semantically void reasoning on beyond-capability tasks, creating risks where plausible-sounding but incorrect derivatio…
EvoBrowseComp: Benchmarking Search Agents on Evolving Knowledge
Yunhan Wang, Jiaan Wang, Lianzhe Huang +2
Search Agents -- large language models augmented with search tools -- have intensified the need for future-proof evaluation benchmarks. Existing benchmarks such as BrowseComp rely…
Investigating Cross-Modal Skill Injection: Scenarios, Methods, and Hyperparameters
Zhiyu Xu, Lean Wang, Yuanxin Liu +5
Vision-Language Models (VLMs) have demonstrated remarkable proficiency in general multi-modal understanding; yet they struggle to efficiently acquire continually evolving domain-sp…
ReflectMT: Internalizing Reflection for Efficient and High-Quality Machine Translation
Kunquan Li, Yingxue Zhang, Fandong Meng +2
Recent years have witnessed growing interest in applying Large Reasoning Models (LRMs) to Machine Translation (MT). While most approaches adopt a 'pre-thinking' paradigm and benefi…
SED-SFT: Selectively Encouraging Diversity in Supervised Fine-Tuning
Yijie Chen, Yijin Liu, Fandong Meng
Supervised Fine-Tuning (SFT) followed by Reinforcement Learning (RL) has emerged as the standard post-training paradigm for large language models (LLMs). However, the conventional…
The Missing Half: Unveiling Training-time Implicit Safety Risks Beyond Deployment
Zhexin Zhang, Yida Lu, Junfeng Fang +8
Safety risks of AI models have been widely studied at deployment time, such as jailbreak attacks that elicit harmful outputs. In contrast, safety risks emerging during training rem…
Figure It Out: Improve the Frontier of Reasoning with Executable Visual States
Meiqi Chen, Fandong Meng, Jie Zhou
Complex reasoning problems often involve implicit spatial and geometric relationships that are not explicitly encoded in text. While recent reasoning models perform well across man…
HGMEM: Hypergraph-based Working Memory to Improve Multi-step RAG for Long-Context Complex Relational Modeling
Chulun Zhou, Chunkang Zhang, Guoxin Yu +4
Multi-step retrieval-augmented generation (RAG) has become a widely adopted strategy for enhancing large language models (LLMs) on tasks that demand global comprehension and intens…
Continuous Autoregressive Language Models
Chenze Shao, Darren Li, Fandong Meng +1
The efficiency of large language models (LLMs) is fundamentally limited by their sequential, token-by-token generation process. We argue that overcoming this bottleneck requires a…
Think Natively: Unlocking Multilingual Reasoning with Consistency-Enhanced Reinforcement Learning
Xue Zhang, Yunlong Liang, Fandong Meng +5
Large Reasoning Models (LRMs) have achieved remarkable performance on complex reasoning tasks by adopting the ``think-then-answer'' paradigm, which enhances both accuracy and inter…
CM-Align: Consistency-based Multilingual Alignment for Large Language Models
Xue Zhang, Yunlong Liang, Fandong Meng +4
Current large language models (LLMs) generally show a significant performance gap in alignment between English and other languages. To bridge this gap, existing research typically…
GRAM-R: Self-Training Generative Foundation Reward Models for Reward Reasoning
Chenglong Wang, Yongyu Mu, Hang Zhou +10
Significant progress in reward modeling over recent years has been driven by a paradigm shift from task-specific designs towards generalist reward models. Despite this trend, devel…
RewardAnything: Generalizable Principle-Following Reward Models
Zhuohao Yu, Jiali Zeng, Weizheng Gu +7
Reward Models, essential for guiding Large Language Model optimization, are typically trained on fixed preference datasets, resulting in rigid alignment to single, implicit prefere…
Dense Retrievers Can Fail on Simple Queries: Revealing The Granularity Dilemma of Embeddings
Liyan Xu, Zhenlin Su, Mo Yu +3
This work stems from an observed limitation of text encoders: embeddings may not be able to recognize fine-grained entities or events within encoded semantics, resulting in failed…
Less, but Better: Efficient Multilingual Expansion for LLMs via Layer-wise Mixture-of-Experts
Xue Zhang, Yunlong Liang, Fandong Meng +4
Continually expanding new languages for existing large language models (LLMs) is a promising yet challenging approach to building powerful multilingual LLMs. The biggest challenge…
SlangDIT: Benchmarking LLMs in Interpretative Slang Translation
Yunlong Liang, Fandong Meng, Jiaan Wang +1
The challenge of slang translation lies in capturing context-dependent semantic extensions, as slang terms often convey meanings beyond their literal interpretation. While slang de…
THOR-MoE: Hierarchical Task-Guided and Context-Responsive Routing for Neural Machine Translation
Yunlong Liang, Fandong Meng, Jie Zhou
The sparse Mixture-of-Experts (MoE) has achieved significant progress for neural machine translation (NMT). However, there exist two limitations in current MoE solutions which may…
ExTrans: Multilingual Deep Reasoning Translation via Exemplar-Enhanced Reinforcement Learning
Jiaan Wang, Fandong Meng, Jie Zhou
In recent years, the emergence of large reasoning models (LRMs), such as OpenAI-o1 and DeepSeek-R1, has shown impressive capabilities in complex problems, e.g., mathematics and cod…
An Empirical Study of Many-to-Many Summarization with Large Language Models
Jiaan Wang, Fandong Meng, Zengkui Sun +5
Many-to-many summarization (M2MS) aims to process documents in any language and generate the corresponding summaries also in any language. Recently, large language models (LLMs) ha…