Large Language Model - 標籤 - DataSci Ocean

Large Language Model

2026

RLHF 讓模型變無聊？揭秘「典型性偏差」與 Verbalized Sampling 如何喚醒 LLM 潛在的創造力 01-03

ACL 2025 | 超越 Self-Consistency！CER 論文解析：無需訓練，靠「過程自信度」激發 Llama 3 與 DeepSeek 推理潛能 01-03

2025

拒絕 AI 一本正經胡說八道！DeepConf 論文解析：如何利用「信心分數」優化推理並大幅節省 Token? 12-28

不微調、不靠 GPT-4！微軟 rStar 如何透過 MCTS 讓 LLaMA2-7B 數學能力暴增 5 倍？ (ICLR 2025) 12-10

你的 Base Model 其實比你想的更聰明！揭秘如何透過 MCMC 採樣讓模型超越 RL 12-02

AI 的下一步：Agentic Context Engineering，讓你的模型學會思考與進化 10-27

AI 如何邊工作邊學習？深入解析《Dynamic Cheatsheet》的自我提升之道 10-15

讓 AI 學會自我除錯！深入了解 SQL-of-Thought 如何打造更精準的 Text-to-SQL 10-12

告別固定的 Top-k！Adaptive-k 如何在不增加延遲下，為 RAG 動態選出最佳上下文 09-22

不只學「對」的，更要學「錯」的？深入解析 SENSE 如何用強弱 LLM 打造頂尖 Text-to-SQL 模型 (ACL 2024) 09-15

[論文介紹] MIRIX: Multi-Agent Memory System for LLM-Based Agents 07-17

[論文介紹] Decoupled Planning and Execution: A Hierarchical Reasoning Framework for Deep Search 07-06

[論文介紹] AutoMind: Adaptive Knowledgeable Agent for Automated Data Science 06-28

[論文介紹] Augmenting LLM Reasoning with Dynamic Notes Writing for Complex QA 06-28

[論文介紹] OctoTools: An Agentic Framework with Extensible Tools for Complex Reasoning 06-02

[論文介紹] Steering Large Language Models Between Code Execution and Textual Reasoning 05-25

[論文介紹] Pre-Act: Multi-Step Planning and Reasoning Improves Acting in LLM Agents 05-20

[論文介紹] MemGPT: Towards LLMs as Operating Systems 05-18

[論文介紹] Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memory 05-13

[論文介紹] PLAN-AND-ACT: Improving Planning of Agents for Long-Horizon Tasks 05-11

1
2