17 papers
UnlearningSoup: Is Repeated Tuning Necessary for Large Language Model Unlearning?
Puning Yang, Qizhou Wang, Junchi Yu +2
Large language models trained on vast corpora inherently risk memorizing harmful content that may later re-emerge in their outputs. To mitigate this issue, existing unlearning meth…
Ripple-Pivot Search: Active Parallel Decoding for Diffusion Large Language Models
Yushi Ye, Xu Chen, Haoyun Jiang +9
Diffusion Large Language Models (dLLMs) have emerged as a competitive alternative to autoregressive language models, offering the potential for substantially faster inference throu…
LiteFrame: Efficient Vision Encoders Unlock Frame Scaling in Video LLMs
Jihwan Kim, Nikhil Parthasarathy, Danfeng Qin +5
The fundamental challenge in scaling Video Large Language Models (Video LLMs) to long-form video lies in managing the explosion of visual-token context length. Existing strategies…
Hidden Human-Like Nature of Machine-Generated Texts: Theory and Detection Enhancement
Chenwang Wu, Yiu-ming Cheung, Bo Han +1
Machine-generated texts (MGTs) produced by large language models (LLMs) are increasingly prevalent across various applications, while their potential misuse in fake news propagatio…
Conditional Equivalence of DPO and RLHF: Implicit Assumption, Failure Modes, and Provable Alignment
Zhiqin Yang, Yonggang Zhang, Wei Xue +3
Direct Preference Optimization (DPO) has emerged as a popular alternative to Reinforcement Learning from Human Feedback (RLHF), offering theoretical equivalence with simpler implem…
What Do Evolutionary Coding Agents Evolve?
Nico Pelleriti, Sree Harsha Nelaturu, Zhanke Zhou +4
Recent work pairs LLMs with evolutionary search to iteratively generate, modify, and select code using task-specific feedback. These systems have produced strong results in mathema…
Rethinking How to Remember: Beyond Atomic Facts in Lifelong LLM Agent Memory
Jingwei Sun, Jianing Zhu, Jiangchao Yao +2
To enable reliable long-term interaction, LLM agents require a memory system that can faithfully store, efficiently retrieve, and deeply reason over accumulated dialogue history. M…
Roll Out and Roll Back: Diffusion LLMs are Their Own Efficiency Teachers
Fanqin Zeng, Feng Hong, Geng Yu +6
Diffusion Large Language Models (DLLMs) promise fast parallel generation, yet open-source DLLMs still face a severe quality-speed trade-off: accelerating decoding by revealing mult…
Distinguishable Deletion: Unifying Knowledge Erasure and Refusal for Large Language Model Unlearning
Puning Yang, Junchi Yu, Qizhou Wang +3
Mitigating sensitive and harmful outputs is fundamental to ensuring safe deployment of LLMs. Existing approaches typically follow two paradigms: Knowledge Deletion (KD), which eras…
Reflective Context Learning: Studying the Optimization Primitives of Context Space
Nikita Vassilyev, William Berrios, Ruowang Zhang +3
Generally capable agents must learn from experience in ways that generalize across tasks and environments. The fundamental problems of learning, including credit assignment, overfi…
AEGIS: Adversarial Target-Guided Retention-Data-Free Robust Concept Erasure from Diffusion Models
Fengpeng Li, Kemou Li, Qizhou Wang +2
Concept erasure helps stop diffusion models (DMs) from generating harmful content; but current methods face robustness retention trade off. Robustness means the model fine-tuned by…
LLM Unlearning with LLM Beliefs
Kemou Li, Qizhou Wang, Yue Wang +4
Large language models trained on vast corpora inherently risk memorizing sensitive or harmful content, which may later resurface in their outputs. Prevailing unlearning methods gen…
EEPO: Exploration-Enhanced Policy Optimization via Sample-Then-Forget
Liang Chen, Xueting Han, Qizhou Wang +4
Balancing exploration and exploitation remains a central challenge in reinforcement learning with verifiable rewards (RLVR) for large language models (LLMs). Current RLVR methods o…
Exploring Criteria of Loss Reweighting to Enhance LLM Unlearning
Puning Yang, Qizhou Wang, Zhuo Huang +3
Loss reweighting has shown significant benefits for machine unlearning with large language models (LLMs). However, their exact functionalities are left unclear and the optimal stra…
LONGER: Scaling Up Long Sequence Modeling in Industrial Recommenders
Zheng Chai, Qin Ren, Xijun Xiao +14
Modeling ultra-long user behavior sequences is critical for capturing both long- and short-term preferences in industrial recommender systems. Existing solutions typically rely on…
GRU: Mitigating the Trade-off between Unlearning and Retention for LLMs
Yue Wang, Qizhou Wang, Feng Liu +4
Large language model (LLM) unlearning has demonstrated its essential role in removing privacy and copyright-related responses, crucial for their legal and safe applications. Howeve…
Rethinking LLM Unlearning Objectives: A Gradient Perspective and Go Beyond
Qizhou Wang, Jin Peng Zhou, Zhanke Zhou +3
Large language models (LLMs) should undergo rigorous audits to identify potential risks, such as copyright and privacy infringements. Once these risks emerge, timely updates are cr…