Alert button
Picture for Zexuan Zhong

Zexuan Zhong

Alert button

REST: Retrieval-Based Speculative Decoding

Nov 14, 2023
Zhenyu He, Zexuan Zhong, Tianle Cai, Jason D Lee, Di He

Viaarxiv icon

Poisoning Retrieval Corpora by Injecting Adversarial Passages

Oct 29, 2023
Zexuan Zhong, Ziqing Huang, Alexander Wettig, Danqi Chen

Viaarxiv icon

Privacy Implications of Retrieval-Based Language Models

May 24, 2023
Yangsibo Huang, Samyak Gupta, Zexuan Zhong, Kai Li, Danqi Chen

Figure 1 for Privacy Implications of Retrieval-Based Language Models
Figure 2 for Privacy Implications of Retrieval-Based Language Models
Figure 3 for Privacy Implications of Retrieval-Based Language Models
Figure 4 for Privacy Implications of Retrieval-Based Language Models
Viaarxiv icon

MQuAKE: Assessing Knowledge Editing in Language Models via Multi-Hop Questions

May 24, 2023
Zexuan Zhong, Zhengxuan Wu, Christopher D. Manning, Christopher Potts, Danqi Chen

Figure 1 for MQuAKE: Assessing Knowledge Editing in Language Models via Multi-Hop Questions
Figure 2 for MQuAKE: Assessing Knowledge Editing in Language Models via Multi-Hop Questions
Figure 3 for MQuAKE: Assessing Knowledge Editing in Language Models via Multi-Hop Questions
Figure 4 for MQuAKE: Assessing Knowledge Editing in Language Models via Multi-Hop Questions
Viaarxiv icon

$k$NN-Adapter: Efficient Domain Adaptation for Black-Box Language Models

Feb 21, 2023
Yangsibo Huang, Daogao Liu, Zexuan Zhong, Weijia Shi, Yin Tat Lee

Figure 1 for $k$NN-Adapter: Efficient Domain Adaptation for Black-Box Language Models
Figure 2 for $k$NN-Adapter: Efficient Domain Adaptation for Black-Box Language Models
Figure 3 for $k$NN-Adapter: Efficient Domain Adaptation for Black-Box Language Models
Figure 4 for $k$NN-Adapter: Efficient Domain Adaptation for Black-Box Language Models
Viaarxiv icon

Training Language Models with Memory Augmentation

May 25, 2022
Zexuan Zhong, Tao Lei, Danqi Chen

Figure 1 for Training Language Models with Memory Augmentation
Figure 2 for Training Language Models with Memory Augmentation
Figure 3 for Training Language Models with Memory Augmentation
Figure 4 for Training Language Models with Memory Augmentation
Viaarxiv icon

Recovering Private Text in Federated Learning of Language Models

May 17, 2022
Samyak Gupta, Yangsibo Huang, Zexuan Zhong, Tianyu Gao, Kai Li, Danqi Chen

Figure 1 for Recovering Private Text in Federated Learning of Language Models
Figure 2 for Recovering Private Text in Federated Learning of Language Models
Figure 3 for Recovering Private Text in Federated Learning of Language Models
Figure 4 for Recovering Private Text in Federated Learning of Language Models
Viaarxiv icon

Structured Pruning Learns Compact and Accurate Models

Apr 01, 2022
Mengzhou Xia, Zexuan Zhong, Danqi Chen

Figure 1 for Structured Pruning Learns Compact and Accurate Models
Figure 2 for Structured Pruning Learns Compact and Accurate Models
Figure 3 for Structured Pruning Learns Compact and Accurate Models
Figure 4 for Structured Pruning Learns Compact and Accurate Models
Viaarxiv icon

Should You Mask 15% in Masked Language Modeling?

Feb 16, 2022
Alexander Wettig, Tianyu Gao, Zexuan Zhong, Danqi Chen

Figure 1 for Should You Mask 15% in Masked Language Modeling?
Figure 2 for Should You Mask 15% in Masked Language Modeling?
Figure 3 for Should You Mask 15% in Masked Language Modeling?
Figure 4 for Should You Mask 15% in Masked Language Modeling?
Viaarxiv icon

Simple Entity-Centric Questions Challenge Dense Retrievers

Sep 26, 2021
Christopher Sciavolino, Zexuan Zhong, Jinhyuk Lee, Danqi Chen

Figure 1 for Simple Entity-Centric Questions Challenge Dense Retrievers
Figure 2 for Simple Entity-Centric Questions Challenge Dense Retrievers
Figure 3 for Simple Entity-Centric Questions Challenge Dense Retrievers
Figure 4 for Simple Entity-Centric Questions Challenge Dense Retrievers
Viaarxiv icon