Elamur Research
☰
home / papers / 2023 / ToolkenGPT: Augmenting Frozen Language Models with Massive Tools via Tool Embeddings

ToolkenGPT: Augmenting Frozen Language Models with Massive Tools via Tool Embeddings

Written for the Research library (claude-opus-5-5) from the paper text, in our own words: the license "arXiv — non-exclusive license to distribute" does not allow redistributing the text itself. Original — arxiv.org/pdf/2305.11554

Summary

Equipping large language models with external tools usually means choosing between two flawed options. Full fine-tuning is expensive and fixes the tool set in advance. In-context prompting hits context-window limits, so it can show only a few tools and examples. ToolkenGPT keeps the LLM frozen and gives each tool its own token, a "toolken." Its embedding is added to the output head and trained on demonstration data. When the model predicts a toolken during normal generation, it moves into a prompted tool mode. There it fills in the arguments, the tool runs, and the result goes back into the text. Tools can be added or removed by changing the toolken set.

The authors test this on:

  • numerical reasoning with math operators (FuncQA),
  • knowledge-base QA with hundreds of relation APIs (KAMEL),
  • embodied planning in VirtualHome.

ToolkenGPT beats prompting baselines such as Chain-of-Thought and ReAct. It comes close to LoRA fine-tuning at a small fraction of the cost. This makes large, changeable tool sets practical for frozen models.

Key points

  • Only the toolken embeddings are trained, and the LLM's weights are never updated. Training costs roughly as much as inference, and parameters are kept separate for each tool, so tools are plug-and-play.
  • Scale of tool sets:
    • over 200 relation APIs in knowledge-based QA;
    • 58 grounded actions and objects for VirtualHome plan generation.
  • Comparison on FuncQA with LLaMA-7B:
    • ToolkenGPT reached 0.55 one-hop accuracy after 2 minutes on one RTX3090.
    • LoRA fine-tuning scored 0.62 but needed 40 minutes on 8×A100.
  • Ablation on FuncQA with LLaMA-30B:
    • ToolkenGPT scored 0.73 one-hop and 0.15 multi-hop.
    • ReAct with an added tool mode scored 0.60 and 0.07.
    • This suggests the embeddings mainly help the model decide when to call a tool and which one.
  • KAMEL test with 30 tools, using 10, 20 or 40 examples per tool:
    • Supervised training data rose from 0.56 to 0.95 accuracy.
    • Synthetic data rose from 0.36 to 0.52.
    • More data helps in both cases, and real supervised data clearly works better.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Other

Cited by

110 papers of our library cite it (of the 140673 whose bibliographies we read); 9 of them have a page here, 101 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.2305.11554 · arXiv 2305.11554

@misc{hao2023toolkengpt,
  title = {{ToolkenGPT: Augmenting Frozen Language Models with Massive Tools via Tool Embeddings}},
  author = {Hao, Shibo and Liu, Tianyang and Wang, Zhen and Hu, Zhiting},
  year = {2023},
  doi = {10.48550/arxiv.2305.11554},
  eprint = {2305.11554},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2305.11554}
}

References

73 references, 27 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Razvan Azamfirei, Sapna R Kudchadkar, and James Fackler. Large language models and the perils of their hallucinations. Critical Care, 27(1):1–2, 2023. find
  2. Michael Bommarito II and Daniel Martin Katz. Gpt takes the bar exam. arXiv preprint arXiv:2212.14402, 2022. — [arXiv:2212.14402](https://arxiv.org/abs/2212.14402) find
  3. Sebastian Borgeaud, Arthur Mensch, Jordan Hoffmann, Trevor Cai, Eliza Rutherford, Katie Millican, George Bm Van Den Driessche, Jean-Baptiste Lespiau, Bogdan Damoc, Aidan Clark, et al. Improving language models by retrieving from trillions of tokens. In International conference on machine learning, p… find
  4. Anthony Brohan, Yevgen Chebotar, Chelsea Finn, Karol Hausman, Alexander Herzog, Daniel Ho, Julian Ibarz, Alex Irpan, Eric Jang, Ryan Julian, et al. Do as i can, not as i say: Grounding language in robotic affordances. In Conference on Robot Learning, pages 287–318. PMLR, 2023. find our card
  5. Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners. Advances in neural information processing systems, 33:1877–1901, 2020. find
  6. Sébastien Bubeck, Varun Chandrasekaran, Ronen Eldan, Johannes Gehrke, Eric Horvitz, Ece Kamar, Peter Lee, Yin Tat Lee, Yuanzhi Li, Scott Lundberg, et al. Sparks of artificial general intelligence: Early experiments with gpt-4. arXiv preprint arXiv:2303.12712, 2023. find our card
  7. Harrison Chase. LangChain, 10 2022. URL https://github.com/hwchase17/langchain. ↗
  8. Wenhu Chen, Xueguang Ma, Xinyi Wang, and William W Cohen. Program of thoughts prompting: Disentangling computation from reasoning for numerical reasoning tasks. arXiv preprint arXiv:2211.12588, 2022. find our card
  9. Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts, Paul Barham, Hyung Won Chung, Charles Sutton, Sebastian Gehrmann, et al. Palm: Scaling language modeling with pathways. arXiv preprint arXiv:2204.02311, 2022. find
  10. Karl Cobbe, Vineet Kosaraju, Mohammad Bavarian, Mark Chen, Heewoo Jun, Lukasz Kaiser, Matthias Plappert, Jerry Tworek, Jacob Hilton, Reiichiro Nakano, et al. Training verifiers to solve math word problems. arXiv preprint arXiv:2110.14168, 2021. find our card
  11. Ning Ding, Yujia Qin, Guang Yang, Fuchao Wei, Zonghan Yang, Yusheng Su, Shengding Hu, Yulin Chen, Chi-Min Chan, Weize Chen, et al. Delta tuning: A comprehensive study of parameter efficient methods for pre-trained language models. arXiv preprint arXiv:2203.06904, 2022. — [arXiv:2203.06904](https://arxiv.org/abs/2203.06904) find
  12. Iddo Drori, Sarah Zhang, Reece Shuttleworth, Leonard Tang, Albert Lu, Elizabeth Ke, Kevin Liu, Linda Chen, Sunny Tran, Newman Cheng, et al. A neural network solves, explains, and generates university math problems by program synthesis and few-shot learning at human level. Proceedings of the National… find
  13. Dheeru Dua, Shivanshu Gupta, Sameer Singh, and Matt Gardner. Successive prompting for decomposing complex questions. arXiv preprint arXiv:2212.04092, 2022. find
  14. Tyna Eloundou, Sam Manning, Pamela Mishkin, and Daniel Rock. Gpts are gpts: An early look at the labor market impact potential of large language models. arXiv preprint arXiv:2303.10130, 2023. — [arXiv:2303.10130](https://arxiv.org/abs/2303.10130) find
  15. Jacqueline Fagard, Lauriane Rat-Fischer, Rana Esseily, Eszter Somogyi, and JK O’Regan. What does it take for an infant to learn how to use a tool by observation? Frontiers in psychology, 7: 267, 2016. find
  16. Bin Fu, Yunqi Qiu, Chengguang Tang, Yang Li, Haiyang Yu, and Jian Sun. A survey on complex question answering over knowledge base: Recent advances and challenges. arXiv preprint arXiv:2007.13069, 2020. find
  17. Luyu Gao, Aman Madaan, Shuyan Zhou, Uri Alon, Pengfei Liu, Yiming Yang, Jamie Callan, and Graham Neubig. Pal: Program-aided language models. arXiv preprint arXiv:2211.10435, 2022. find
  18. Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat, and Mingwei Chang. Retrieval augmented language model pre-training. In International conference on machine learning, pages 3929–3938. PMLR, 2020. find
  19. Shibo Hao, Yi Gu, Haodi Ma, Joshua Jiahua Hong, Zhen Wang, Daisy Zhe Wang, and Zhiting Hu. Reasoning with language model is planning with world model. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, 2023. find
  20. Shibo Hao, Bowen Tan, Kaiwen Tang, Bin Ni, Xiyan Shao, Hengzhe Zhang, Eric Xing, and Zhiting Hu. BertNet: Harvesting knowledge graphs with arbitrary relations from pretrained language models. In Findings of the Association for Computational Linguistics: ACL 2023, pages 5000– 5015, Toronto, Canada, J… find
  21. Joy He-Yueya, Gabriel Poesia, Rose E Wang, and Noah D Goodman. Solving math word problems by combining language models with symbolic solvers. arXiv preprint arXiv:2304.09102, 2023. find
  22. Neil Houlsby, Andrei Giurgiu, Stanislaw Jastrzebski, Bruna Morrone, Quentin De Laroussilhe, Andrea Gesmundo, Mona Attariyan, and Sylvain Gelly. Parameter-efficient transfer learning for nlp. In International Conference on Machine Learning, pages 2790–2799. PMLR, 2019. find
  23. Edward J Hu, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, et al. Lora: Low-rank adaptation of large language models. In International Conference on Learning Representations, 2021. find
  24. Zhiting Hu and Eric P Xing. Toward a ’standard model’ of machine learning. Harvard Data Science Review, 2022. find
  25. Wenlong Huang, Pieter Abbeel, Deepak Pathak, and Igor Mordatch. Language models as zero-shot planners: Extracting actionable knowledge for embodied agents. In International Conference on Machine Learning, pages 9118–9147. PMLR, 2022. find our card
  26. Wenlong Huang, Fei Xia, Ted Xiao, Harris Chan, Jacky Liang, Pete Florence, Andy Zeng, Jonathan Tompson, Igor Mordatch, Yevgen Chebotar, et al. Inner monologue: Embodied reasoning through planning with language models. arXiv preprint arXiv:2207.05608, 2022. find
  27. Wenlong Huang, Fei Xia, Dhruv Shah, Danny Driess, Andy Zeng, Yao Lu, Pete Florence, Igor Mordatch, Sergey Levine, Karol Hausman, et al. Grounded decoding: Guiding text generation with grounded models for robot control. arXiv preprint arXiv:2303.00855, 2023. — [arXiv:2303.00855](https://arxiv.org/abs/2303.00855) find
  28. Ziwei Ji, Nayeon Lee, Rita Frieske, Tiezheng Yu, Dan Su, Yan Xu, Etsuko Ishii, Ye Jin Bang, Andrea Madotto, and Pascale Fung. Survey of hallucination in natural language generation. ACM Computing Surveys, 55(12):1–38, 2023. find our card
  29. Qiao Jin, Yifan Yang, Qingyu Chen, and Zhiyong Lu. Genegpt: Augmenting large language models with domain tools for improved access to biomedical information. ArXiv, 2023. find
  30. Jan-Christoph Kalo and Leandra Fichtel. Kamel: Knowledge analysis with multitoken entities in language models. In Proceedings of the Conference on Automated Knowledge Base Construction, 2022. find
  31. Rabeeh Karimi Mahabadi, James Henderson, and Sebastian Ruder. Compacter: Efficient low-rank hypercomplex adapter layers. Advances in Neural Information Processing Systems, 34:1022–1035, 2021. find
  32. Tushar Khot, Harsh Trivedi, Matthew Finlayson, Yao Fu, Kyle Richardson, Peter Clark, and Ashish Sabharwal. Decomposed prompting: A modular approach for solving complex tasks. arXiv preprint arXiv:2210.02406, 2022. find
  33. Yann LeCun. A path towards autonomous machine intelligence version 0.9. 2, 2022-06-27. Open Review, 62, 2022. find
  34. Brian Lester, Rami Al-Rfou, and Noah Constant. The power of scale for parameter-efficient prompt tuning. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 3045–3059, 2021. find
  35. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, et al. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Processing Systems, 33:9459–9474, 2020. find
  36. Minghao Li, Feifan Song, Bowen Yu, Haiyang Yu, Zhoujun Li, Fei Huang, and Yongbin Li. Api-bank: A benchmark for tool-augmented llms. arXiv preprint arXiv:2304.08244, 2023. — [arXiv:2304.08244](https://arxiv.org/abs/2304.08244) find
  37. Shuang Li, Xavier Puig, Chris Paxton, Yilun Du, Clinton Wang, Linxi Fan, Tao Chen, De-An Huang, Ekin Akyürek, Anima Anandkumar, et al. Pre-trained language models for interactive decision-making. Advances in Neural Information Processing Systems, 35:31199–31212, 2022. find
  38. Xiang Lisa Li and Percy Liang. Prefix-tuning: Optimizing continuous prompts for generation. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 4582–4597,… find
  39. Yaobo Liang, Chenfei Wu, Ting Song, Wenshan Wu, Yan Xia, Yu Liu, Yang Ou, Shuai Lu, Lei Ji, Shaoguang Mao, et al. Taskmatrix. ai: Completing tasks by connecting foundation models with millions of apis. arXiv preprint arXiv:2303.16434, 2023. — [arXiv:2303.16434](https://arxiv.org/abs/2303.16434) find
  40. Bo Liu, Yuqian Jiang, Xiaohan Zhang, Qiang Liu, Shiqi Zhang, Joydeep Biswas, and Peter Stone. Llm+ p: Empowering large language models with optimal planning proficiency. arXiv preprint arXiv:2304.11477, 2023. find our card
  41. Pengfei Liu, Weizhe Yuan, Jinlan Fu, Zhengbao Jiang, Hiroaki Hayashi, and Graham Neubig. Pre-train, prompt, and predict: A systematic survey of prompting methods in natural language processing. ACM Computing Surveys, 55(9):1–35, 2023. find our card
  42. Xiao Liu, Kaixuan Ji, Yicheng Fu, Weng Lam Tam, Zhengxiao Du, Zhilin Yang, and Jie Tang. P-tuning v2: Prompt tuning can be comparable to fine-tuning universally across scales and tasks. arXiv preprint arXiv:2110.07602, 2021. find our card
  43. Pan Lu, Baolin Peng, Hao Cheng, Michel Galley, Kai-Wei Chang, Ying Nian Wu, Song-Chun Zhu, and Jianfeng Gao. Chameleon: Plug-and-play compositional reasoning with large language models. arXiv preprint arXiv:2304.09842, 2023. find our card
  44. Qing Lyu, Shreya Havaldar, Adam Stein, Li Zhang, Delip Rao, Eric Wong, Marianna Apidianaki, and Chris Callison-Burch. Faithful chain-of-thought reasoning. arXiv preprint arXiv:2301.13379, 2023. find our card
  45. Grégoire Mialon, Roberto Dessì, Maria Lomeli, Christoforos Nalmpantis, Ram Pasunuru, Roberta Raileanu, Baptiste Rozière, Timo Schick, Jane Dwivedi-Yu, Asli Celikyilmaz, et al. Augmented language models: a survey. arXiv preprint arXiv:2302.07842, 2023. find our card
  46. Reiichiro Nakano, Jacob Hilton, Suchir Balaji, Jeff Wu, Long Ouyang, Christina Kim, Christopher Hesse, Shantanu Jain, Vineet Kosaraju, William Saunders, et al. Webgpt: Browser-assisted question-answering with human feedback. arXiv preprint arXiv:2112.09332, 2021. find our card
  47. OpenAI. Gpt-4 technical report, 2023. find our card
  48. Batu Ozturkler, Nikolay Malkin, Zhen Wang, and Nebojsa Jojic. Thinksum: Probabilistic reasoning over sets using large language models. arXiv preprint arXiv:2210.01293, 2022. find
  49. Bhargavi Paranjape, Scott Lundberg, Sameer Singh, Hannaneh Hajishirzi, Luke Zettlemoyer, and Marco Tulio Ribeiro. Art: Automatic multi-step reasoning and tool-use for large language models. arXiv preprint arXiv:2303.09014, 2023. find
  50. Aaron Parisi, Yao Zhao, and Noah Fiedel. Talm: Tool augmented language models. arXiv preprint arXiv:2205.12255, 2022. find our card
  51. Fabio Petroni, Tim Rocktäschel, Patrick Lewis, Anton Bakhtin, Yuxiang Wu, Alexander H Miller, and Sebastian Riedel. Language models as knowledge bases? arXiv preprint arXiv:1909.01066, 2019. find our card
  52. Xavier Puig, Kevin Ra, Marko Boben, Jiaman Li, Tingwu Wang, Sanja Fidler, and Antonio Torralba. Virtualhome: Simulating household activities via programs. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 8494–8502, 2018. find
  53. Yujia Qin, Shengding Hu, Yankai Lin, Weize Chen, Ning Ding, Ganqu Cui, Zheni Zeng, Yufei Huang, Chaojun Xiao, Chi Han, Yi Ren Fung, Yusheng Su, Huadong Wang, Cheng Qian, Runchu Tian, Kunlun Zhu, Shi Liang, Xingyu Shen, Bokai Xu, Zhen Zhang, Yining Ye, Bo Li, Ziwei Tang, Jing Yi, Yu Zhu, Zhenning Dai… find our card
  54. Nils Reimers and Iryna Gurevych. Sentence-bert: Sentence embeddings using siamese bert-networks. arXiv preprint arXiv:1908.10084, 2019. find
  55. Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Eric Michael Smith, Y-Lan Boureau, et al. Recipes for building an open-domain chatbot. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main… find our card
  56. Timo Schick, Jane Dwivedi-Yu, Roberto Dessì, Roberta Raileanu, Maria Lomeli, Luke Zettlemoyer, Nicola Cancedda, and Thomas Scialom. Toolformer: Language models can teach themselves to use tools. arXiv preprint arXiv:2302.04761, 2023. find our card
  57. Yongliang Shen, Kaitao Song, Xu Tan, Dongsheng Li, Weiming Lu, and Yueting Zhuang. Hugginggpt: Solving ai tasks with chatgpt and its friends in huggingface. arXiv preprint arXiv:2303.17580, 2023. find our card
  58. Kurt Shuster, Spencer Poff, Moya Chen, Douwe Kiela, and Jason Weston. Retrieval augmentation reduces hallucination in conversation. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 3784–3803, 2021. find our card
  59. Ishika Singh, Valts Blukis, Arsalan Mousavian, Ankit Goyal, Danfei Xu, Jonathan Tremblay, Dieter Fox, Jesse Thomason, and Animesh Garg. Progprompt: Generating situated robot task plans using large language models. arXiv preprint arXiv:2209.11302, 2022. find our card
  60. Alon Talmor and Jonathan Berant. The web as a knowledge-base for answering complex questions. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 641–651, 2018. find
  61. Romal Thoppilan, Daniel De Freitas, Jamie Hall, Noam Shazeer, Apoorv Kulshreshtha, Heng-Tze Cheng, Alicia Jin, Taylor Bos, Leslie Baker, Yu Du, et al. Lamda: Language models for dialog applications. arXiv preprint arXiv:2201.08239, 2022. find our card
  62. Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971, 2023. find
  63. Yizhong Wang, Yeganeh Kordi, Swaroop Mishra, Alisa Liu, Noah A Smith, Daniel Khashabi, and Hannaneh Hajishirzi. Self-instruct: Aligning language model with self generated instructions. arXiv preprint arXiv:2212.10560, 2022. find
  64. Zhen Wang, Rameswar Panda, Leonid Karlinsky, Rogerio Feris, Huan Sun, and Yoon Kim. Multitask prompt tuning enables parameter-efficient transfer learning. In The Eleventh International Conference on Learning Representations, 2023. find
  65. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Ed Chi, Quoc Le, and Denny Zhou. Chain of thought prompting elicits reasoning in large language models. arXiv preprint arXiv:2201.11903, 2022. find our card
  66. Jiannan Xiang, Tianhua Tao, Yi Gu, Tianmin Shu, Zirui Wang, Zichao Yang, and Zhiting Hu. Language Models Meet World Models: Embodied Experiences Enhance Language Models. NeurIPS, 2023. find
  67. Yaqi Xie, Chen Yu, Tongyao Zhu, Jinbin Bai, Ze Gong, and Harold Soh. Translating natural language to planning goals with large-language models. arXiv preprint arXiv:2302.05128, 2023. — [arXiv:2302.05128](https://arxiv.org/abs/2302.05128) find
  68. Sherry Yang, Ofir Nachum, Yilun Du, Jason Wei, Pieter Abbeel, and Dale Schuurmans. Foundation models for decision making: Problems, methods, and opportunities. arXiv preprint arXiv:2303.04129, 2023. find
  69. Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, and Yuan Cao. React: Synergizing reasoning and acting in language models. arXiv preprint arXiv:2210.03629, 2022. find our card
  70. Wenhao Yu, Chenguang Zhu, Zaitang Li, Zhiting Hu, Qingyun Wang, Heng Ji, and Meng Jiang. A survey of knowledge-enhanced text generation. ACM Computing Surveys, 54(11s):1–38, 2022. find
  71. Elad Ben Zaken, Yoav Goldberg, and Shauli Ravfogel. Bitfit: Simple parameter-efficient fine-tuning for transformer-based masked language-models. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 1–9, 2022. find our card
  72. Yuheng Zha, Yichi Yang, Ruichen Li, and Zhiting Hu. AlignScore: Evaluating factual consistency with a unified alignment function. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 11328–11348, Toronto, Canada, July 2023. Associa… find our card
  73. Yuheng Zha, Yichi Yang, Ruichen Li, and Zhiting Hu. Text alignment is an efficient unified model for massive nlp tasks. NeurIPS, 2023. find