Výsledky vyhledávání

Report

Beyond Examples: High-level Automated Reasoning Paradigm in In-Context Learning via MCTS

Autor: Wu, Jinyang, Feng, Mingkuan, Zhang, Shuai, Che, Feihu, Wen, Zengqi, Tao, Jianhua

In-context Learning (ICL) enables large language models (LLMs) to tackle downstream tasks through sophisticated prompting and high-quality demonstrations. However, this traditional ICL paradigm shows limitations when facing complex mathematical reaso

Externí odkaz: http://arxiv.org/abs/2411.18478

Zobrazit plný text záznamu

Report

Pandora's Box or Aladdin's Lamp: A Comprehensive Analysis Revealing the Role of RAG Noise in Large Language Models

Autor: Wu, Jinyang, Che, Feihu, Zhang, Chuyuan, Tao, Jianhua, Zhang, Shuai, Shao, Pengpeng

Retrieval-Augmented Generation (RAG) has emerged as a crucial method for addressing hallucinations in large language models (LLMs). While recent research has extended RAG models to complex noisy scenarios, these explorations often confine themselves

Externí odkaz: http://arxiv.org/abs/2408.13533

Zobrazit plný text záznamu

Report

Regularity of viscosity solutions of the $\sigma_k$-Yamabe-type Problem for $k>n/2$

Autor: Wu, Jinyang

We study the regularity of Lipschitz viscosity solutions to the $\sigma_k$ Yamabe problem in the negative cone case. If either $k=n$ or the manifold is conformally flat and $k>n/2$, we prove that all Lipschitz viscosity solutions are smooth away from

Externí odkaz: http://arxiv.org/abs/2407.08300

Zobrazit plný text záznamu

Report

Can large language models understand uncommon meanings of common words?

Autor: Wu, Jinyang, Che, Feihu, Zheng, Xinxin, Zhang, Shuai, Jin, Ruihan, Nie, Shuai, Shao, Pengpeng, Tao, Jianhua

Large language models (LLMs) like ChatGPT have shown significant advancements across diverse natural language understanding (NLU) tasks, including intelligent dialogue and autonomous agents. Yet, lacking widely acknowledged testing mechanisms, answer

Externí odkaz: http://arxiv.org/abs/2405.05741

Zobrazit plný text záznamu

Report

KS-LLM: Knowledge Selection of Large Language Models with Evidence Document for Question Answering

Autor: Zheng, Xinxin, Che, Feihu, Wu, Jinyang, Zhang, Shuai, Nie, Shuai, Liu, Kang, Tao, Jianhua

Large language models (LLMs) suffer from the hallucination problem and face significant challenges when applied to knowledge-intensive tasks. A promising approach is to leverage evidence documents as extra supporting knowledge, which can be obtained

Externí odkaz: http://arxiv.org/abs/2404.15660

Zobrazit plný text záznamu

Report

A new renormalized volume type invariant

Autor: Wu, Jinyang

Publikováno v: Pacific J. Math. 334 (2025) 43-57

In this paper, we define a new conformal invariant on complete non-compact hyperbolic surfaces that can be conformally compactified to bounded domains in $\mathbb{C}$. We study and compute this invariant up to one-connected surfaces. Our results give

Externí odkaz: http://arxiv.org/abs/2309.12268

Zobrazit plný text záznamu