4 citations · 19 across the 32 of their papers we have counts for
25 papers · 1 filter
From Input to Output: A Flexible Agent for Dual-End Interpretation of Sparse Autoencoder Features
Dewen Liu, Zixuan Li, Jonathan Pan +4
Sparse autoencoders (SAEs) are an important tool for mechanistic interpretability, but interpreting their many features remains challenging. Existing methods characterize input-sid…
From Base Rollouts to RL Reasoning: A Budgeted Search Perspective
Wenhe Sun, Cunxiang Wang, Zijun Yao +1
Reinforcement learning with verifiable rewards (RLVR) improves language-model reasoning, but how these gains relate to inference-time decoding and search remains unclear. Does RL c…
Where Steering Signals Come From: Activation Source Selection in Activation Steering
Jiaran Ye, Lingxu Ran, Zijun Yao +5
Activation steering controls language models by adding vectors or features to hidden states at inference time, but the upstream source of these steering signals is often treated as…
SAEVerbalizer: Generating Explanations for Sparse Autoencoder Features via Representation Verbalization
Weihan Meng, Hongzhu Guo, Yi Jing +5
Sparse autoencoders (SAEs) are proposed to extract numerous features from large language model (LLM) representations, yet explaining these features still relies primarily on extern…
WildReward: Learning Reward Models from In-the-Wild Human Interactions
Hao Peng, Yunjia Qi, Xiaozhi Wang +3
Reward models (RMs) are crucial for the training of large language models (LLMs), yet they typically rely on large-scale human-annotated preference pairs. With the widespread deplo…
MM-THEBench: Do Reasoning MLLMs Think Reasonably?
Zhidian Huang, Zijun Yao, Ji Qi +7
Recent advances in multimodal large language models (MLLMs) mark a shift from non-thinking models to post-trained reasoning models capable of solving complex problems through think…
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…
Auxiliary Metrics Help Decoding Skill Neurons in the Wild
Yixiu Zhao, Xiaozhi Wang, Zijun Yao +2
Large language models (LLMs) exhibit remarkable capabilities across a wide range of tasks, yet their internal mechanisms remain largely opaque. In this paper, we introduce a simple…
GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models
5 Team, Aohan Zeng, Xin Lv +167
We present GLM-4.5, an open-source Mixture-of-Experts (MoE) large language model with 355B total parameters and 32B activated parameters, featuring a hybrid reasoning method that s…
Are Reasoning Models More Prone to Hallucination?
Zijun Yao, Yantao Liu, Yanxu Chen +5
Recently evolved large reasoning models (LRMs) show powerful performance in solving complex tasks with long chain-of-thought (CoT) reasoning capability. As these LRMs are mostly de…
Agentic Reward Modeling: Integrating Human Preferences with Verifiable Correctness Signals for Reliable Reward Systems
Hao Peng, Yunjia Qi, Xiaozhi Wang +4
Reward models (RMs) are crucial for the training and inference-time scaling up of large language models (LLMs). However, existing reward models primarily focus on human preferences…
PairJudge RM: Perform Best-of-N Sampling with Knockout Tournament
Yantao Liu, Zijun Yao, Rui Min +3
Best-of-N (BoN) sampling, a common strategy for test-time scaling of Large Language Models (LLMs), relies on reward models to select the best candidate solution from multiple gener…
LinguaLens: Towards Interpreting Linguistic Mechanisms of Large Language Models via Sparse Auto-Encoder
Yi Jing, Zijun Yao, Hongzhu Guo +4
Large language models (LLMs) demonstrate exceptional performance on tasks requiring complex linguistic abilities, such as reference disambiguation and metaphor recognition/generati…
AtomR: Atomic Operator-Empowered Large Language Models for Heterogeneous Knowledge Reasoning
Amy Xin, Jinxin Liu, Zijun Yao +4
Despite the outstanding capabilities of large language models (LLMs), knowledge-intensive reasoning still remains a challenging task due to LLMs' limitations in compositional reaso…
Pre-training Distillation for Large Language Models: A Design Space Exploration
Hao Peng, Xin Lv, Yushi Bai +4
Knowledge distillation (KD) aims to transfer knowledge from a large teacher model to a smaller student model. Previous work applying KD in the field of large language models (LLMs)…
LLMAEL: Large Language Models are Good Context Augmenters for Entity Linking
Amy Xin, Yunjia Qi, Zijun Yao +5
Specialized entity linking (EL) models are well-trained at mapping mentions to unique knowledge base (KB) entities according to a given context. However, specialized EL models stru…
Aligning Teacher with Student Preferences for Tailored Training Data Generation
Yantao Liu, Zhao Zhang, Zijun Yao +3
Large Language Models (LLMs) have shown significant promise as copilots in various tasks. Local deployment of LLMs on edge devices is necessary when handling privacy-sensitive data…
SeaKR: Self-aware Knowledge Retrieval for Adaptive Retrieval Augmented Generation
Zijun Yao, Weijian Qi, Liangming Pan +5
This paper introduces Self-aware Knowledge Retrieval (SeaKR), a novel adaptive RAG model that extracts self-aware uncertainty of LLMs from their internal states. SeaKR activates re…
DICE: Detecting In-distribution Contamination in LLM's Fine-tuning Phase for Math Reasoning
Shangqing Tu, Kejian Zhu, Yushi Bai +3
The advancement of large language models (LLMs) relies on evaluation using public benchmarks, but data contamination can lead to overestimated performance. Previous researches focu…
SoAy: A Solution-based LLM API-using Methodology for Academic Information Seeking
Yuanchun Wang, Jifan Yu, Zijun Yao +13
Applying large language models (LLMs) for academic API usage shows promise in reducing researchers' academic information seeking efforts. However, current LLM API-using methods str…
Transferable and Efficient Non-Factual Content Detection via Probe Training with Offline Consistency Checking
Xiaokang Zhang, Zijun Yao, Jing Zhang +4
Detecting non-factual content is a longstanding goal to increase the trustworthiness of large language models (LLMs) generations. Current factuality probes, trained using humananno…
Untangle the KNOT: Interweaving Conflicting Knowledge and Reasoning Skills in Large Language Models
Yantao Liu, Zijun Yao, Xin Lv +5
Providing knowledge documents for large language models (LLMs) has emerged as a promising solution to update the static knowledge inherent in their parameters. However, knowledge i…
Evaluating Generative Language Models in Information Extraction as Subjective Question Correction
Yuchen Fan, Yantao Liu, Zijun Yao +3
Modern Large Language Models (LLMs) have showcased remarkable prowess in various tasks necessitating sophisticated cognitive behaviors. Nevertheless, a paradoxical performance disc…
A Cause-Effect Look at Alleviating Hallucination of Knowledge-grounded Dialogue Generation
Jifan Yu, Xiaohan Zhang, Yifan Xu +5
Empowered by the large-scale pretrained language models, existing dialogue systems have demonstrated impressive performance conducting fluent and natural-sounding conversations. Ho…