Elamur Research⌕
☰
home / papers / 2024 / AgentBoard: An Analytical Evaluation Board of Multi-turn LLM Agents

AgentBoard: An Analytical Evaluation Board of Multi-turn LLM Agents

Paper · 2024 · Chang Ma, Junlei Zhang, Zhihao Zhu et al.

Summary

Most benchmarks for LLM agents judge models only by whether they finish a task. This hides differences when nearly every model fails and reveals little about why. Many benchmarks also rely on single-turn or fully observable problems. AgentBoard (NeurIPS 2024, Datasets and Benchmarks track) addresses this with 1013 text-based environments spread over nine tasks in four areas: embodied AI, games, web, and tools. Humans checked each environment to confirm it requires several rounds of interaction under partial observability. Subgoals were defined or hand-annotated for every sample, which allows a progress rate that credits partial advancement. An open-source toolkit with a Wandb-based panel lets users break down results by difficulty, sub-skill, interaction length, action validity (grounding), and full trajectories. The authors used a simple reflex agent that maps the goal plus history to the next action, and tested both proprietary and open-weight models. GPT-4 was best overall, and Llama3 and DeepSeek LLM were the strongest open models. Progress rate exposed gains that success rate missed. Weak multi-turn ability set most open models apart. The resource supports diagnosing agent weaknesses, not just ranking systems.

Key points

  • Tasks: AlfWorld, ScienceWorld, BabyAI (embodied); Jericho, PDDL (game); WebShop, WebArena (web); Tool Query, Tool Operation (tool). Together they contain 1013 environments.
  • Interaction is modeled as a POMDP. The baseline agent conditions on the goal and on a memory of past observations and actions.
  • Six sub-skills are scored: memory, planning, world modeling, self-reflection, grounding, and spatial navigation. GPT-4 leads on all of them, and open-weight models show uneven gaps.
  • The authors cite Ruan et al. (2024), who found that AgentBoard scores correlate with reasoning and coding benchmarks, while AgentBench scores track the knowledge test MMLU.
  • Limitations: progress rate depends on human-written subgoals, since current LLMs cannot generate them reliably, and evaluation is restricted to simulated environments. Code and data are available at github.com/hkust-nlp/AgentBoard.

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, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 works

Cite

Cite as Chang Ma, Junlei Zhang, Zhihao Zhu, Cheng Yang, Yujiu Yang, Yaohui Jin, Zhenzhong Lan, Lingpeng Kong and Junxian He. 2024. AgentBoard: An Analytical Evaluation Board of Multi-turn LLM Agents. arXiv:2401.13178.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.2401.13178 · arXiv 2401.13178

@misc{ma2024agentboard,
  title = {{AgentBoard: An Analytical Evaluation Board of Multi-turn LLM Agents}},
  author = {Chang Ma, Junlei Zhang, Zhihao Zhu, Cheng Yang, Yujiu Yang, Yaohui Jin, Zhenzhong Lan, Lingpeng Kong and Junxian He},
  year = {2024},
  doi = {10.48550/arxiv.2401.13178},
  eprint = {2401.13178},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2401.13178}
}

References

62 references, 33 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Ahn, M., Brohan, A., Brown, N., Chebotar, Y., Cortes, O., David, B., Finn, C., Fu, C., Gopalakrishnan, K., Hausman, K., et al. (2022). Do as i can, not as i say: Grounding language in robotic affordances. ArXiv preprint, abs/2204.01691.
  2. Anthropic (2023). Introducing claude. find
  3. Chase, H. (2022). Langchain. find
  4. Chevalier-Boisvert, M., Bahdanau, D., Lahlou, S., Willems, L., Saharia, C., Nguyen, T. H., and Bengio, Y. (2019). Babyai: A platform to study the sample efficiency of grounded language learning. In 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 20…
  5. Chiang, W.-L., Li, Z., Lin, Z., Sheng, Y., Wu, Z., Zhang, H., Zheng, L., Zhuang, S., Zhuang, Y., Gonzalez, J. E., et al. (2023). Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality. See https://vicuna.lmsys. org (accessed 14 April 2023). ↗
  6. Cobbe, K., Kosaraju, V., Bavarian, M., Chen, M., Jun, H., Kaiser, L., Plappert, M., Tworek, J., Hilton, J., Nakano, R., Hesse, C., and Schulman, J. (2021). Training verifiers to solve math word problems. arXiv preprint arXiv:2110.14168.
  7. DeepSeek-AI, :, Bi, X., Chen, D., Chen, G., Chen, S., Dai, D., Deng, C., Ding, H., Dong, K., Du, Q., Fu, Z., Gao, H., Gao, K., Gao, W., Ge, R., Guan, K., Guo, D., Guo, J., Hao, G., Hao, Z., He, Y., Hu, W., Huang, P., Li, E., Li, G., Li, J., Li, Y., Li, Y. K., Liang, W., Lin, F., Liu, A. X., Liu, B.,… find
  8. Deng, X., Gu, Y., Zheng, B., Chen, S., Stevens, S., Wang, B., Sun, H., and Su, Y. (2023). Mind2web: Towards a generalist agent for the web. ArXiv preprint, abs/2306.06070.
  9. Driess, D., Xia, F., Sajjadi, M. S., Lynch, C., Chowdhery, A., Ichter, B., Wahid, A., Tompson, J., Vuong, Q., Yu, T., et al. (2023). Palm-e: An embodied multimodal language model. ArXiv preprint, abs/2303.03378. find
  10. Dziri, N., Lu, X., Sclar, M., Li, X. L., Jian, L., Lin, B. Y., West, P., Bhagavatula, C., Bras, R. L., Hwang, J. D., et al. (2023). Faith and fate: Limits of transformers on compositionality. ArXiv preprint, abs/2305.18654.
  11. Gu, Y., Deng, X., and Su, Y. (2022). Don’t generate, discriminate: A proposal for grounding language models to real-world environments. ArXiv preprint, abs/2212.09736. find
  12. Hausknecht, M. J., Ammanabrolu, P., Côté, M., and Yuan, X. (2020). Interactive fiction games: A colossal adventure. In The Thirty-Fourth AAAI Conference on Artificial Intelligence, AAAI 2020, The Thirty-Second Innovative Applications of Artificial Intelligence Conference, IAAI 2020, The Tenth AAAI S… find
  13. Hong, S., Zheng, X., Chen, J., Cheng, Y., Wang, J., Zhang, C., Wang, Z., Yau, S. K. S., Lin, Z., Zhou, L., et al. (2023). Metagpt: Meta programming for multi-agent collaborative framework. arXiv preprint arXiv:2308.00352.
  14. Jiang, A. Q., Sablayrolles, A., Mensch, A., Bamford, C., Chaplot, D. S., Casas, D. d. l., Bressand, F., Lengyel, G., Lample, G., Saulnier, L., et al. (2023). Mistral 7b. ArXiv preprint, abs/2310.06825. find
  15. Kaplan, J., McCandlish, S., Henighan, T., Brown, T. B., Chess, B., Child, R., Gray, S., Radford, A., Wu, J., and Amodei, D. (2020). Scaling laws for neural language models. arXiv preprint arXiv:2001.08361.
  16. Kinniment, M., Sato, L. J. K., Du, H., Goodrich, B., Hasin, M., Chan, L., Miles, L. H., Lin, T. R., Wijk, H., Burget, J., et al. (2023). Evaluating language-model agents on realistic autonomous tasks. ArXiv preprint, abs/2312.11671. — [arXiv:2312.11671](https://arxiv.org/abs/2312.11671) find
  17. Kwon, W., Li, Z., Zhuang, S., Sheng, Y., Zheng, L., Yu, C. H., Gonzalez, J., Zhang, H., and Stoica, I. (2023). Efficient memory management for large language model serving with pagedattention. In Proceedings of the 29th Symposium on Operating Systems Principles, pages 611–626.
  18. Lanham, T., Chen, A., Radhakrishnan, A., Steiner, B., Denison, C., Hernandez, D., Li, D., Durmus, E., Hubinger, E., Kernion, J., et al. (2023). Measuring faithfulness in chain-of-thought reasoning. ArXiv preprint, abs/2307.13702.
  19. LeCun, Y. (2022). A path towards autonomous machine intelligence version 0.9. 2, 2022-06-27. Open Review, 62. find
  20. Li, M., Song, F., Yu, B., Yu, H., Li, Z., Huang, F., and Li, Y. (2023). Api-bank: A benchmark for tool-augmented llms. ArXiv preprint, abs/2304.08244. — [arXiv:2304.08244](https://arxiv.org/abs/2304.08244) find
  21. Lin, Y.-T. and Chen, Y.-N. (2023). Llm-eval: Unified multi-dimensional automatic evaluation for open-domain conversations with large language models. ArXiv preprint, abs/2305.13711.
  22. Liu, X., Yu, H., Zhang, H., Xu, Y., Lei, X., Lai, H., Gu, Y., Ding, H., Men, K., Yang, K., et al. (2023a). Agentbench: Evaluating llms as agents. ArXiv preprint, abs/2308.03688.
  23. Liu, Z., Yao, W., Zhang, J., Xue, L., Heinecke, S., Murthy, R., Feng, Y., Chen, Z., Niebles, J. C., Arpit, D., et al. (2023b). Bolaa: Benchmarking and orchestrating llm-augmented autonomous agents. ArXiv preprint, abs/2308.05960. find
  24. Madaan, A., Tandon, N., Gupta, P., Hallinan, S., Gao, L., Wiegreffe, S., Alon, U., Dziri, N., Prabhumoye, S., Yang, Y., et al. (2023). Self-refine: Iterative refinement with self-feedback. ArXiv preprint, abs/2303.17651. find
  25. Mialon, G., Fourrier, C., Swift, C., Wolf, T., LeCun, Y., and Scialom, T. (2023). Gaia: a benchmark for general ai assistants. ArXiv preprint, abs/2311.12983.
  26. OpenAI (2022). Introducing chatgpt. find
  27. OpenAI (2023). Gpt-4 technical report. arXiv, pages 2303–08774.
  28. Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., et al. (2022). Training language models to follow instructions with human feedback. Advances in Neural Information Processing Systems, 35:27730–27744.
  29. Pourchot, A. and Sigaud, O. (2019). CEM-RL: combining evolutionary and gradient-based methods for policy search. In 7th International Conference on Learning Representations, ICLR 2019, New Orleans, LA, USA, May 6-9, 2019. OpenReview.net. find
  30. Puterman, M. L. (1990). Markov decision processes. Handbooks in operations research and management science, 2:331–434. find
  31. Qin, Y., Hu, S., Lin, Y., Chen, W., Ding, N., Cui, G., Zeng, Z., Huang, Y., Xiao, C., Han, C., et al. (2023a). Tool learning with foundation models. ArXiv preprint, abs/2304.08354.
  32. Qin, Y., Liang, S., Ye, Y., Zhu, K., Yan, L., Lu, Y., Lin, Y., Cong, X., Tang, X., Qian, B., et al. (2023b). Toolllm: Facilitating large language models to master 16000+ real-world apis. ArXiv preprint, abs/2307.16789.
  33. Reed, S., Zolna, K., Parisotto, E., Colmenarejo, S. G., Novikov, A., Barth-Maron, G., Gimenez, M., Sulsky, Y., Kay, J., Springenberg, J. T., et al. (2022). A generalist agent. ArXiv preprint, abs/2205.06175. find
  34. Richards, T. B. (2023). Significant-gravitas/autogpt: An experimental open-source attempt to make gpt-4 fully autonomous. find
  35. Roziere, B., Gehring, J., Gloeckle, F., Sootla, S., Gat, I., Tan, X. E., Adi, Y., Liu, J., Remez, T., Rapin, J., et al. (2023). Code llama: Open foundation models for code. ArXiv preprint, abs/2308.12950.
  36. Ruan, Y., Maddison, C. J., and Hashimoto, T. (2024). Observational scaling laws and the predictability of language model performance. arXiv preprint arXiv:2405.10938.
  37. Russell, S. and Norvig, P. (2005). Ai a modern approach. Learning, 2(3):4. find
  38. Shi, T., Karpathy, A., Fan, L., Hernandez, J., and Liang, P. (2017a). World of bits: An open-domain platform for web-based agents. In Precup, D. and Teh, Y. W., editors, Proceedings of the 34th International Conference on Machine Learning, ICML 2017, Sydney, NSW, Australia, 6-11 August 2017, volume… find
  39. Shi, T., Karpathy, A., Fan, L., Hernandez, J., and Liang, P. (2017b). World of bits: An open-domain platform for web-based agents. In Precup, D. and Teh, Y. W., editors, Proceedings of the 34th International Conference on Machine Learning, ICML 2017, Sydney, NSW, Australia, 6-11 August 2017, volume… find
  40. Shinn, N., Cassano, F., Gopinath, A., Narasimhan, K. R., and Yao, S. (2023). Reflexion: Language agents with verbal reinforcement learning. In Thirty-seventh Conference on Neural Information Processing Systems.
  41. Shridhar, M., Yuan, X., Côté, M., Bisk, Y., Trischler, A., and Hausknecht, M. J. (2021). Alfworld: Aligning text and embodied environments for interactive learning. In 9th International Conference on Learning Representations, ICLR 2021, Virtual Event, Austria, May 3-7, 2021. OpenReview.net.
  42. Silver, T. and Chitnis, R. (2020). Pddlgym: Gym environments from pddl problems. ArXiv preprint, abs/2002.06432. — [arXiv:2002.06432](https://arxiv.org/abs/2002.06432) find
  43. Song, C. H., Wu, J., Washington, C., Sadler, B. M., Chao, W.-L., and Su, Y. (2023). Llm-planner: Few-shot grounded planning for embodied agents with large language models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 2998–3009.
  44. Touvron, H., Lavril, T., Izacard, G., Martinet, X., Lachaux, M.-A., Lacroix, T., Rozière, B., Goyal, N., Hambro, E., Azhar, F., et al. (2023). Llama: Open and efficient foundation language models. ArXiv preprint, abs/2302.13971.
  45. Vallati, M., Chrpa, L., Grześ, M., McCluskey, T. L., Roberts, M., Sanner, S., et al. (2015). find
  46. The 2014 international planning competition: Progress and trends. Ai Magazine, 36(3):90–98. find
  47. Wang, G., Xie, Y., Jiang, Y., Mandlekar, A., Xiao, C., Zhu, Y., Fan, L., and Anandkumar, A. (2023a). Voyager: An open-ended embodied agent with large language models. ArXiv preprint, abs/2305.16291. find
  48. Wang, R., Jansen, P., Côté, M.-A., and Ammanabrolu, P. (2022). ScienceWorld: Is your agent smarter than a 5th grader? In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 11279–11298, Abu Dhabi, United Arab Emirates. Association for Computational Linguisti… find
  49. Wang, X., Wang, Z., Liu, J., Chen, Y., Yuan, L., Peng, H., and Ji, H. (2023b). Mint: Evaluating llms in multi-turn interaction with tools and language feedback. ArXiv preprint, abs/2309.10691.
  50. Wei, J., Tay, Y., Bommasani, R., Raffel, C., Zoph, B., Borgeaud, S., Yogatama, D., Bosma, M., Zhou, D., Metzler, D., et al. (2022). Emergent abilities of large language models. ArXiv preprint, abs/2206.07682.
  51. Wu, Y., Tang, X., Mitchell, T. M., and Li, Y. (2023). Smartplay: A benchmark for llms as intelligent agents. ArXiv preprint, abs/2310.01557. find
  52. Wu, Z., Han, C., Ding, Z., Weng, Z., Liu, Z., Yao, S., Yu, T., and Kong, L. (2024). Os-copilot: Towards generalist computer agents with self-improvement. arXiv preprint arXiv:2402.07456.
  53. Xie, T., Zhou, F., Cheng, Z., Shi, P., Weng, L., Liu, Y., Hua, T. J., Zhao, J., Liu, Q., Liu, C., et al. (2023). Openagents: An open platform for language agents in the wild. ArXiv preprint, abs/2310.10634.
  54. Xu, B., Liu, X., Shen, H., Han, Z., Li, Y., Yue, M., Peng, Z., Liu, Y., Yao, Z., and Xu, D. (2023a). Gentopia: A collaborative platform for tool-augmented llms. arXiv preprint arXiv:2308.04030. find
  55. Xu, Q., Hong, F., Li, B., Hu, C., Chen, Z., and Zhang, J. (2023b). On the tool manipulation capability of open-source large language models. ArXiv preprint, abs/2305.16504.
  56. Xu, Y., Su, H., Xing, C., Mi, B., Liu, Q., Shi, W., Hui, B., Zhou, F., Liu, Y., Xie, T., et al. (2023c). Lemur: Harmonizing natural language and code for language agents. ArXiv preprint, abs/2310.06830. — [arXiv:2310.06830](https://arxiv.org/abs/2310.06830) find
  57. Yang, Z., Liu, J., Han, Y., Chen, X., Huang, Z., Fu, B., and Yu, G. (2023). Appagent: Multimodal agents as smartphone users. ArXiv preprint, abs/2312.13771.
  58. Yao, S., Chen, H., Yang, J., and Narasimhan, K. (2022). Webshop: Towards scalable real-world web interaction with grounded language agents. Advances in Neural Information Processing Systems, 35:20744–20757.
  59. Yao, S., Zhao, J., Yu, D., Du, N., Shafran, I., Narasimhan, K. R., and Cao, Y. (2023). ReAct: Synergizing reasoning and acting in language models. In The Eleventh International Conference on Learning Representations.
  60. Ye, J., Li, G., Gao, S., Huang, C., Wu, Y., Li, S., Fan, X., Dou, S., Zhang, Q., Gui, T., et al. (2024). Tooleyes: Fine-grained evaluation for tool learning capabilities of large language models in real-world scenarios. ArXiv preprint, abs/2401.00741.
  61. Zheng, B., Gou, B., Kil, J., Sun, H., and Su, Y. (2024). Gpt-4v(ision) is a generalist web agent, if grounded. ArXiv preprint, abs/2401.01614.
  62. Zhou, S., Xu, F. F., Zhu, H., Zhou, X., Lo, R., Sridhar, A., Cheng, X., Bisk, Y., Fried, D., Alon, U., et al. (2023). Webarena: A realistic web environment for building autonomous agents. ArXiv preprint, abs/2307.13854.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/ma-2024-agentboard-analytical-evaluation-board-multi/