15 papers
Which Knowledge Should LLMs Trust? Uncovering a Generalizable Steering Direction Across Knowledge Conflicts
Yibin Lou, Yi Hu, Jianhui Chen +3
Large Language Models (LLMs) frequently encounter conflicts between internal parametric knowledge and external contextual knowledge, yet the mechanisms governing this selection rem…
Knowledge Interaction in Hypernetwork-based Continual Learning: Conflict and Composition
Jiaqi Zheng, Yi Hu, Yadong Mu +1
Continual learning aims to adapt pretrained large language models (LLMs) to evolving data while preserving their pretrained capabilities. A central challenge is catastrophic forget…
Shared Weights, Selected Computations: How Looped Transformers Route What Each Loop Does
Jiaju Wu, Yi Hu, Muhan Zhang
Looped Transformers repeatedly apply the same set of Transformer layers, giving them a recurrent architecture for latent computation. Their strong performance on iterative reasonin…
ContinualSkillBench: Can LLM Agents Truly Evolve Their Capabilities?
Tianyi Guan, Yiding Wang, Haotong Yang +5
Modern agent frameworks equip large language models with external skill libraries to solve complex tasks. However, it remains unclear whether these systems can effectively evolve t…
GABench: A Comprehensive Benchmark for Evaluating LLM Agents on Graph Analysis Tasks
Jiarui Tan, Zhongjian Zhang, YaBo Guo +5
Large language model (LLM) agents are increasingly capable of planning, using tools, and interacting with external environments. They are typically supported by harnesses, which ma…
The Expressivity Boundary of Probabilistic Circuits: A Comparison with Large Language Models
Zhiyu Zhao, Xuejie Liu, Muhan Zhang +1
Probabilistic Circuits (PCs) are deep generative models that support exact and efficient probabilistic inference. Yet in autoregressive language modeling, PCs still lag behind Tran…
SAGE: A Self-Evolving Agentic Graph-Memory Engine for Structure-Aware Associative Memory
Juntong Wang, Haoyue Zhao, guanghui Pan +4
Long-term memory is becoming a central bottleneck for language agents. Exsting RAG and GraphRAG systems largely treat memory graphs as static retrieval middleware, which limits the…
Position: How can Graphs Help Large Language Models?
Xiyuan Wang, Yi Hu, Yanbo Wang +2
With the rapid advancement of large language models (LLMs), classic graph learning tasks have greatly benefited from LLMs, including improved encoding of textual features, more eff…
ProofVerifier: A Scalable, Diversity-Driven Framework for Natural-Language Proof Verification
Haotong Yang, Zitong Wang, Shijia Kang +7
While large language models (LLMs) have achieved strong performance on mathematical problems with verifiable answers, many advanced problems are proof-based and require evaluating…
SubTokenTest: A Practical Benchmark for Real-World Sub-token Understanding
Shuyang Hou, Yi Hu, Muhan Zhang
Recent advancements in large language models (LLMs) have significantly enhanced their reasoning capabilities. However, they continue to struggle with basic character-level tasks, s…
Towards a Mechanistic Understanding of Large Reasoning Models: A Survey of Training, Inference, and Failures
Yi Hu, Jiaqi Gu, Ruxin Wang +6
Reinforcement learning (RL) has catalyzed the emergence of Large Reasoning Models (LRMs) that have pushed reasoning capabilities to new heights. While their performance has garnere…
What Affects the Effective Depth of Large Language Models?
Yi Hu, Cai Zhou, Muhan Zhang
The scaling of large language models (LLMs) emphasizes increasing depth, yet performance gains diminish with added layers. Prior work introduces the concept of "effective depth", a…
The Road Less Traveled: Enhancing Exploration in LLMs via Sequential Sampling
Shijia Kang, Muhan Zhang
Reinforcement learning (RL) has been pivotal in enhancing the reasoning capabilities of large language models (LLMs), but it often suffers from limited exploration and entropy coll…
Coevolutionary Continuous Discrete Diffusion: Make Your Diffusion Language Model a Latent Reasoner
Cai Zhou, Chenxiao Yang, Yi Hu +7
Diffusion language models, especially masked discrete diffusion models, have achieved great success recently. While there are some theoretical and primary empirical results showing…
PHYBench: Holistic Evaluation of Physical Perception and Reasoning in Large Language Models
Shi Qiu, Shaoyang Guo, Zhuo-Yang Song +51
Current benchmarks for evaluating the reasoning capabilities of Large Language Models (LLMs) face significant limitations: task oversimplification, data contamination, and flawed e…