Výsledky vyhledávání - "Chen, Hongzhan"

Report

SocialBench: Sociality Evaluation of Role-Playing Conversational Agents

Autor: Chen, Hongzhan, Chen, Hehong, Yan, Ming, Xu, Wenshen, Gao, Xing, Shen, Weizhou, Quan, Xiaojun, Li, Chenliang, Zhang, Ji, Huang, Fei, Zhou, Jingren

Large language models (LLMs) have advanced the development of various AI conversational agents, including role-playing conversational agents that mimic diverse characters and human behaviors. While prior research has predominantly focused on enhancin

Externí odkaz: http://arxiv.org/abs/2403.13679

Zobrazit plný text záznamu

Report

Small LLMs Are Weak Tool Learners: A Multi-LLM Agent

Autor: Shen, Weizhou, Li, Chenliang, Chen, Hongzhan, Yan, Ming, Quan, Xiaojun, Chen, Hehong, Zhang, Ji, Huang, Fei

Large Language Model (LLM) agents significantly extend the capabilities of standalone LLMs, empowering them to interact with external tools (e.g., APIs, functions) and complete various tasks in a self-directed fashion. The challenge of tool use deman

Externí odkaz: http://arxiv.org/abs/2401.07324

Zobrazit plný text záznamu

Report

Knowledge Distillation of Black-Box Large Language Models

Autor: Chen, Hongzhan, Chen, Ruijun, Yi, Yuqi, Quan, Xiaojun, Li, Chenliang, Yan, Ming, Zhang, Ji

Given the exceptional performance of proprietary large language models (LLMs) like GPT-4, recent research has increasingly focused on boosting the capabilities of smaller models through knowledge distillation (KD) from these powerful yet black-box te

Externí odkaz: http://arxiv.org/abs/2401.07013

Zobrazit plný text záznamu

Report

MCC-KD: Multi-CoT Consistent Knowledge Distillation

Autor: Chen, Hongzhan, Wu, Siyue, Quan, Xiaojun, Wang, Rui, Yan, Ming, Zhang, Ji

Large language models (LLMs) have showcased remarkable capabilities in complex reasoning through chain of thought (CoT) prompting. Recently, there has been a growing interest in transferring these reasoning abilities from LLMs to smaller models. Howe

Externí odkaz: http://arxiv.org/abs/2310.14747

Zobrazit plný text záznamu

Report

AD-KD: Attribution-Driven Knowledge Distillation for Language Model Compression

Autor: Wu, Siyue, Chen, Hongzhan, Quan, Xiaojun, Wang, Qifan, Wang, Rui

Knowledge distillation has attracted a great deal of interest recently to compress pre-trained language models. However, existing knowledge distillation methods suffer from two limitations. First, the student model simply imitates the teacher's behav

Externí odkaz: http://arxiv.org/abs/2305.10010

Zobrazit plný text záznamu