14 papers · 1 filter
Selective Knowledge Edit Reversal via Gated Singular Vector Shrinkage
Weifeng Jiang, Ruirui Chen, Qianren Mao +3
Knowledge editing provides an efficient way to update factual knowledge in large language models. However, malicious edits may introduce safety risks, making it necessary to revers…
E-Orch: Towards Effective, Efficient, and Extensible Agentic Orchestration with Reinforcement Learning
Junnan Liu, Linhao Luo, Thuy-Trang Vu +1
Agentic orchestration enables multiple autonomous agents to solve complex tasks through adaptive decomposition, delegation, and execution. However, existing orchestrators often rel…
Enhancing Multilingual Reasoning via Steerable Model Merging
Zhuoran Li, Rui Xu, Jian Yang +8
Model merging is an effective technique for composing the capabilities of a multilingual model and a reasoning model. It has achieved promising generalization in multilingual reaso…
OpenCompass: A Universal Evaluation Platform for Large Language Models
Maosong Cao, Kai Chen, Haodong Duan +27
In recent years, the field of artificial intelligence has undergone a paradigm shift from task-specific small-scale models to general-purpose large language models (LLMs). With the…
AIPO: Learning to Reason from Active Interaction
Junnan Liu, Linhao Luo, Thuy-Trang Vu +1
Recent advances in LLMs have demonstrated strong reasoning capabilities, largely stimulated by RLVR. However, the exploration of existing RLVR algorithms remains largely constraine…
Rectifying LLM Thought from Lens of Optimization
Junnan Liu, Hongwei Liu, Songyang Zhang +1
Recent advancements in large language models (LLMs) have been driven by their emergent reasoning capabilities, particularly through long chain-of-thought (CoT) prompting, which ena…
ATLAS: A High-Difficulty, Multidisciplinary Benchmark for Frontier Scientific Reasoning
Hongwei Liu, Junnan Liu, Shudong Liu +33
The rapid advancement of Large Language Models (LLMs) has led to performance saturation on many established benchmarks, questioning their ability to distinguish frontier models. Co…
CompassVerifier: A Unified and Robust Verifier for LLMs Evaluation and Outcome Reward
Shudong Liu, Hongwei Liu, Junnan Liu +8
Answer verification is crucial not only for evaluating large language models (LLMs) by matching their unstructured outputs against standard answers, but also serves as the reward m…
When Tools Hurt LLM Reasoning: State-Dependent Belief Revision under External Evidence
Yufeng Zhao, Junnan Liu, Hongwei Liu +4
Tool use is often assumed to monotonically improve reasoning, where external evidence is expected to help when relevant and be ignored when irrelevant. We show that this assumption…
Rethinking Verification for LLM Code Generation: From Generation to Testing
Zihan Ma, Taolin Zhang, Maosong Cao +5
Large language models (LLMs) have recently achieved notable success in code-generation benchmarks such as HumanEval and LiveCodeBench. However, a detailed examination reveals that…
Coding Triangle: How Does Large Language Model Understand Code?
Taolin Zhang, Zihan Ma, Maosong Cao +3
Large language models (LLMs) have achieved remarkable progress in code generation, yet their true programming competence remains underexplored. We introduce the Code Triangle frame…
Deciphering Trajectory-Aided LLM Reasoning: An Optimization Perspective
Junnan Liu, Hongwei Liu, Linchen Xiao +5
We propose a novel framework for comprehending the reasoning capabilities of large language models (LLMs) through the perspective of meta-learning. By conceptualizing reasoning tra…
SituatedThinker: Grounding LLM Reasoning with Real-World through Situated Thinking
Junnan Liu, Linhao Luo, Thuy-Trang Vu +1
Recent advances in large language models (LLMs) demonstrate their impressive reasoning capabilities. However, the reasoning confined to internal parametric space limits LLMs' acces…
Exploring the Limit of Outcome Reward for Learning Mathematical Reasoning
Chengqi Lyu, Songyang Gao, Yuzhe Gu +14
Reasoning abilities, especially those for solving complex math problems, are crucial components of general intelligence. Recent advances by proprietary companies, such as o-series…