11 papers
Agentic Unlearning: When LLM Agent Meets Machine Unlearning
Bin Wang, Fan Wang, Pingping Wang +5
In this paper, we introduce \textbf{agentic unlearning} which removes specified information from both model parameters and persistent memory in agents with closed-loop interaction.…
TrajAD: Trajectory Anomaly Detection for Trustworthy LLM Agents
Yibing Liu, Chong Zhang, Zhongyi Han +5
We address the problem of runtime trajectory anomaly detection, a critical capability for enabling trustworthy LLM agents. Current safety measures predominantly focus on static inp…
Towards Automated Kernel Generation in the Era of LLMs
Yang Yu, Peiyu Zang, Chi Hsu Tsai +11
The performance of modern AI systems is fundamentally constrained by the quality of their underlying GPU kernels, which translate high-level algorithmic semantics into low-level ha…
Boosting RL-Based Visual Reasoning with Selective Adversarial Entropy Intervention
Yang Yu, Zhuangzhuang Chen, Lanqing Li +1
Recently, reinforcement learning (RL) has become a common choice in enhancing the reasoning capabilities of vision-language models (VLMs). Considering existing RL-based finetuning…
Pass@k Metric for RLVR: A Diagnostic Tool of Exploration, But Not an Objective
Yang Yu
The ability of Large Language Models (LLMs) to perform complex, multi-step reasoning is a central focus of modern AI research. To evaluate and enhance this capability, the pass@k m…
Reinforcement Learning with Anticipation: A Hierarchical Approach for Long-Horizon Tasks
Yang Yu
Solving long-horizon goal-conditioned tasks remains a significant challenge in reinforcement learning (RL). Hierarchical reinforcement learning (HRL) addresses this by decomposing…
LLM Data Selection and Utilization via Dynamic Bi-level Optimization
Yang Yu, Kai Han, Hang Zhou +4
While large-scale training data is fundamental for developing capable large language models (LLMs), strategically selecting high-quality data has emerged as a critical approach to…
CCI4.0: A Bilingual Pretraining Dataset for Enhancing Reasoning in Large Language Models
Guang Liu, Liangdong Wang, Jijie Li +6
We introduce CCI4.0, a large-scale bilingual pre-training dataset engineered for superior data quality and diverse human-like reasoning trajectory. CCI4.0 occupies roughly TB…
DTVM: Revolutionizing Smart Contract Execution with Determinism and Compatibility
Wei Zhou, Xiong Xu, Changzheng Wei +22
We introduce the DeTerministic Virtual Machine (DTVM) Stack, a next-generation smart contract execution framework designed to address critical performance, determinism, and ecosyst…
NeoRL-2: Near Real-World Benchmarks for Offline Reinforcement Learning with Extended Realistic Scenarios
Songyi Gao, Zuolin Tu, Rong-Jun Qin +3
Offline reinforcement learning (RL) aims to learn from historical data without requiring (costly) access to the environment. To facilitate offline RL research, we previously introd…
Efficient Recurrent Off-Policy RL Requires a Context-Encoder-Specific Learning Rate
Fan-Ming Luo, Zuolin Tu, Zefang Huang +1
Real-world decision-making tasks are usually partially observable Markov decision processes (POMDPs), where the state is not fully observable. Recent progress has demonstrated that…