metadata
license: apache-2.0
datasets:
- ai2lumos/lumos_complex_qa_plan_iterative
language:
- en
tags:
- language-agent
- question-answering
- reasoning
- grounding
馃獎 Lumos: Language Agents with Unified Formats, Modular Design, and Open-Source LLMs
馃寪[Website] 馃摑[Paper] 馃[Data] 馃[Model]
We introduce 馃獎Lumos, Language Agents with Unified Formats, Modular Design, and Open-Source LLMs. Lumos unifies a suite of complex interactive tasks and achieves competitive performance with GPT-4/3.5-based and larger open-source agents.
Lumos has following features:
- 馃З Modular Architecture:
- Lumos consists of planning, grounding, and execution modules built based on LLAMA-2-7B.
- 馃實 Diverse Training Data:
- Lumos is trained with ~40K high-quality annotations from ground-truth reasoning steps in existing benchmarks with GPT-4.
- 馃殌 Competitive Performance:
- 馃殌 Lumos outperforms GPT-4/3.5-based agents on complex QA and web agent tasks, and larger open agents on maths tasks.
- 馃殌 Lumos performs better than open agent baseline formulations including chain-of-thoughts and unmodularized training.
- 馃殌 Lumos surpasses larger open LLM agents and domain-specific agents on an unseen task, WebShop.
Model Overview
lumos_complex_qa_plan_iterative is a planning module checkpoint finetuned on complex QA task in Lumos-Iterative (Lumos-I) formulation.
The training annotation is shown below:
| Training Data | Number |
|---|---|
lumos_complex_qa_plan_iterative |
19409 |
Citation
If you find this work is relevant with your research, please feel free to cite our work!
@article{yin2023lumos,
title={Lumos: Towards Language Agents that are Unified, Modular, and Open Source},
author={Yin, Da and Brahman, Faeze and Ravichander, Abhilasha and Chandu, Khyathi and Chang, Kai-Wei and Choi, Yejin and Lin, Bill Yuchen},
year={2023}
}