<Tool> ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs

<Tool> ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs

최근(2023.07)에 나온 논문을 읽어보고 간단히 정리했습니다. 혹시 부족하거나 잘못된 내용이 있다면 댓글 부탁드립니다 ️ usechatgpt init success data construction과 model training/evaluation을 위한 프레임워크, ToolLLM tool 사용을 위해 제작된 instruction tuning dataset, ToolBench LLaMA 모델을 ToolBench에 fine-tuning한 ToolLLaMA 배경 직전 paper review에서 설명한 바와 같이, 최근 LLM의 성장세가 눈부심에도 불구하고 여러 high-level task에 대해 아쉬운 성능을 보일 때가 많다는 한계에 대해 여러 지적이 나오고 있습니다. 당연한 이야기이지만 LLM을 학습..

원문링크 : <Tool> ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs

 등록된 다른 글

[BOJ] 11286 : 절댓값 힙 [우선순위 큐](Python)

[BOJ] 11286 : 절댓값 힙 [우선순위 큐](Python)

<Decompose> Self-Discover: Large Language Models Self-Compose Reasoning Structures (2024.02)

<Decompose> Self-Discover: Large Language Models Self-Compose Reasoning Structures (2024.02)

[프로그래머스] 콜라 문제(Python)

[프로그래머스] 콜라 문제(Python)

<Alignment> RLAIF: Scaling Reinforcement Learning from Human Feedback with AI Feedback

<Alignment> RLAIF: Scaling Reinforcement Learning from Human Feedback with AI Feedback

[비전공자 취업 회고록] 영어영문학 전공자가 데이터 사이언티스트가 되기까지는 1년이 걸렸다(feat. 비전공자인데 개발자 해도 될까요?)

[비전공자 취업 회고록] 영어영문학 전공자가 데이터 사이언티스트가 되기까지는 1년이 걸렸다(feat. 비전공자인데 개발자 해도 될까요?)

<Prompting, Decomposition> Least-to-Most Prompting Enables Complex Reasoning in Large Language Models (2023.04)

<Prompting, Decomposition> Least-to-Most Prompting Enables Complex Reasoning in Large Language Models (2023.04)

<Prompt, Agent> [SPP] Unleashing Cognitive Synergy in Large Language Models: A Task-Solving Agent through Multi-Persona Self-Collaboration (2023.07)

<Prompt, Agent> [SPP] Unleashing Cognitive Synergy in Large Language Models: A Task-Solving Agent through Multi-Persona Self-Collaboration (2023.07)

[Short Paper Review] Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes

[Short Paper Review] Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes