Agentlm 70b

🚀 AgentLM-70B

AgentTuning is the first attempt to instruction-tune large language models (LLMs) using interaction trajectories across multiple agent tasks. Evaluation results show that AgentTuning endows LLMs with agent capabilities, which can generalize well on unseen agent tasks while maintaining good general language abilities. We have open-sourced the AgentInstruct dataset and AgentLM.

🤗 [Dataset] • 💻 [Github Repo] • 📌 [Project Page] • 📃 [Paper]

✨ Features

• AgentTuning enables LLMs to have agent capabilities with robust generalization on unseen agent tasks.
• Maintains good general language abilities.
• Open-sourced the AgentInstruct dataset and AgentLM.

📚 Documentation

Models

AgentLM models are produced by mixed training on the AgentInstruct dataset and the ShareGPT dataset from Llama - 2 - chat models.

The models follow the conversation format of Llama-2-chat, with the system prompt fixed as

You are a helpful, respectful and honest assistant.

7B, 13B, and 70B models are available on the Huggingface model hub.

📄 License

Citation

If you find our work useful, please consider citing AgentTuning:

@misc{zeng2023agenttuning,
      title={AgentTuning: Enabling Generalized Agent Abilities for LLMs}, 
      author={Aohan Zeng and Mingdao Liu and Rui Lu and Bowen Wang and Xiao Liu and Yuxiao Dong and Jie Tang},
      year={2023},
      eprint={2310.12823},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}