REVIEW 5 major objections 5 minor 45 references
TUMS: Enhancing Tool-use Abilities of LLMs with Multi-structure Handlers
T0 review · 5 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash
Pith's one-line read Generating tool parameters with per-tool prompt structures raises LLM tool-use accuracy by 19.6% on easy and 50.6% on hard ToolQA benchmark questions.
desk verdict Solid easy-question gains on ToolQA, but the paper's central attribution to multi-structure handlers is undercut by hand-tuned, benchmark-specific design choices. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central mechanism is the multi-structure handler set inside the subtask processor: three few-shot prompt structures, direct, parallel, and serial, each matched to tools by a heuristic rule based on parameter count and usage difficulty. The parallel structure handles tools whose parameters are many but independent, and the serial structure handles tools whose parameters must be discovered sequentially, with each stage's output feeding the next. The framework's own ablation demonstrates that the mechanism, not the surrounding pipeline, carries the accuracy gain: replacing all three structures with only the direct one lowers the correct rate by roughly 19.9 percentage points to below the ReAct baseline. The intent recognizer contributes mainly efficiency, reducing the average number of model queries needed per correct answer.
What would settle it
Take a held-out toolset, assign each tool's handler structure automatically from its schema or without looking at the test questions, run TUMS against ReAct, and check whether the advantage survives. A second check: give the one-structure variant the same few-shot examples inside a single combined prompt, and if it then matches TUMS, the gain comes from exemplar count or prompt length rather than from the structural split.
Extended reading notes
Core claim
The discovery, on the paper's own terms, is that parameter generation should be moved from the tool level to the parameter level. Because tools differ in how many parameters they need and how hard they are to invoke, a single generation structure produces missing, irrelevant, or wrong parameters, so the authors assign each tool to one of three handler structures based on its difficulty. Direct Generation handles simple few-parameter calls; Parallel Generation splits multi-condition filters into categories such as time, space, and object, generates each condition separately, and integrates them; Serial Generation builds complex calls like SQL queries step by step, from query framework to column-to-value mapping to final synthesis. Wrapped in an intent recognizer that narrows the toolset and a decomposer that iterates subtask-tool pairs against execution results, the design yields the reported ToolQA gains, and a preference-based variant that favors shorter solving paths reaches a 40.13% overall correct rate against ReAct's 29.93%.
Load-bearing premise
The load-bearing premise is that the hand-made assignment of each tool to a direct, parallel, or serial handler, which the paper itself labels as based on 'subjective complexity' and which was made with knowledge of the benchmark's tools, is what produces the reported gains; if those structure choices do not transfer to other tool sets, the improvement could vanish.
Editorial extensions
If this is right
- Other tool-augmented agents could adopt the same parameter-level split without retraining, because each tool's generation structure is simply a few-shot prompt and the gain should transfer to any sufficiently capable instruction-following model.
- The one-structure ablation implies that intent recognition and task decomposition alone will not beat ReAct on ToolQA; the handler split is what pushes accuracy above the baselines.
- Because the intent recognizer lowers the average number of model queries per correct answer, a pipeline with more planning stages can still be cheaper per success than a simpler one.
- The processor is explicitly expandable to more structures, so the three-handler design is a starting point rather than a ceiling for covering additional tool types.
- Preference-based hints that encode "the shorter solving path is the better path" push the overall correct rate from 29.93% for ReAct to 40.13% for TUMS-PRE, pointing to planning priors as a further lever.
Reading between the lines
- If the framework is to generalize, the structure choice itself needs automation: a model that selects direct, parallel, or serial generation from a tool's own schema, tested on tools not seen during design, would establish transferability.
- A matched-exemplar ablation, in which the one-structure variant receives the same few-shot examples inside a single combined prompt, would separate the effect of structure choice from the effect of extra exemplars, which the reported comparison leaves entangled.
- The parameter-level idea is not confined to question answering: any agent that builds arguments of varying size and coupling, such as REST API calls with optional fields or code interpreters, could use the same direct-parallel-serial split.
- TUMS-PRE's preference for short tool paths suggests that a cheap planning prior can substitute for expensive search over tool sequences, a hypothesis worth testing on agents that currently rely on tree or graph search for tool planning.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes TUMS, a four-module framework for LLM tool use consisting of an intent recognizer, a task decomposer, a subtask processor with multiple parameter-generation handlers (Direct, Parallel, and Serial), and an executor. The core claim is that transforming tool-level processing into parameter-level processing via multi-structure handlers improves tool-use accuracy. Experiments on ToolQA with Qwen1.5-72B report an average 19.6% improvement over the previous best baseline on easy questions and 50.6% on hard questions, along with ablations of the handler structures and of the intent recognizer, and an efficiency analysis.
Significance. If the attribution claimed in the paper were established, the idea of assigning structure-specific parameter-generation procedures according to tool complexity would be a useful and easily adoptable contribution to tool-augmented LLMs. The paper deserves credit for a clear modular design, a published prompt suite, an ablation of the multi-structure mechanism, and a cost analysis. However, the current evidence is in-sample: the handler assignments and the preference-based hints are hand-designed for ToolQA, and the hard-question average is dominated by a single dataset, so the general significance of the framework is not yet demonstrated.
major comments (5)
- [Section 3.3 and Appendix 6.2] The assignment of each tool to the Direct, Parallel, or Serial handler structure is made by the authors' 'subjective complexity', and Appendix 6.2 introduces preference-based hints that are manually written for specific ToolQA questions. Because the paper provides no automatic structure-selection mechanism and no transfer experiment to unseen tools or datasets, the experiments do not establish that parameter-level multi-structure handlers, rather than benchmark-specific prompt engineering, are the cause of the reported improvements.
- [Tables 1 and 2, Section 4.2] The abstract's claim of 'almost an overall enhancement' is not supported on the hard benchmark: in Table 2, TUMS degrades relative to ReAct on Flight-hard (11.0 to 6.0) and Airbnb-hard (10.0 to 7.0), while the 50.6% average is heavily driven by Yelp-hard (14.0 to 36.0). The authors themselves note 'unstable performance in hard questions of table datasets'; the discussion should analyze per-dataset variability instead of relying on the average, and should report variance or significance testing.
- [Section 4.1 and following paragraph] The Qwen and CoT baselines are not given access to any tools, so their near-zero scores on tool-dependent datasets are not informative for a tool-use comparison. The meaningful tool-using baselines are ReAct and Chameleon, and against them TUMS is not uniformly better (for example, it is worse than ReAct on Flight-hard). The main comparison should be reframed accordingly, and paired or resampling-based significance tests should be reported.
- [Figure 3 and Section 4.2] The ablation TUMS vs. TUMS-OS is described as holding exemplars fixed while removing the multi-structure handlers, but the paper does not specify how a Direct-only prompt is constructed for tools that originally used Parallel or Serial exemplars, and no variance or significance information is given. The ablation therefore does not conclusively isolate the multi-structure handlers from the accompanying tool-specific exemplars, which is the central attribution of the paper.
- [Appendix 6.2] The TUMS-PRE variant introduces preference-based hints that are manually written for specific hard questions after observing the framework's bottlenecks, as illustrated by the Airbnb question in Appendix 6.2. The reported 40.13% correct rate for TUMS-PRE is therefore a benchmark-specific result and should not be used as evidence for the general effectiveness of the framework; it should be clearly labeled as an in-sample variant or removed from the contribution claims.
minor comments (5)
- [Tables 1 and 2] The arrows in the TUMS rows are not defined; the reader cannot tell whether the percentages compare TUMS to ReAct specifically or to the best baseline in each column, and the dash on Coffee-easy should be explained.
- [Figure 4] The x-axis labels 'On Easy Questions' and 'On Hard Questions' are confusing because both the total cost and the average cost appear on the same plot; the figure should be clarified or split into separate panels.
- [References] References 29 and 30 are identical citations of Shi et al., 'Learning to use tools via cooperative and interactive agents'; the duplicate should be removed and the in-text citations reconciled.
- [Appendix 6.1] The paper states that the code has been open-sourced in GitHub, but no repository URL or identifier is provided; please add one for reproducibility.
- [Throughout] There are several typos and style inconsistencies, including 'Prompt F or' in Figure 5, 'such markable performance improvement' in Section 4.2, and the first-person 'Acknowledgments' section; these should be corrected in a final revision.
Circularity Check
No circularity: the reported gains are empirical evaluation of a disclosed framework, not a derivation from its own assumptions.
full rationale
TUMS is an empirical framework, so there is no mathematical derivation chain in which a claimed output reduces to an input by construction. The multi-structure handlers are introduced as an explicit design choice (Section 3.3: 'we use a divide-and-conquer strategy that heuristically decomposes the parameter-generation process based on the subjective complexity of the tool'), and the three structures are specified before evaluation rather than fitted to the results. The main comparison against baselines and the ablation (TUMS vs. TUMS-OS vs. TUMS-NIR) are executed on ToolQA, with TUMS-OS described as using the same exemplars as TUMS but only a Direct Generation Structure, which is a reasonable control for the multi-structure contribution. The preference-based hint in Appendix 6.2 is disclosed as a separate variant, TUMS-PRE, not as the source of the headline 19.6% and 50.6% averages. The defensible weaknesses are generalization and validity concerns: the manual assignment of tools to handler structures is done with knowledge of ToolQA, and no automatic structure-selection mechanism is provided. Those are benchmark-specificity issues, not circularity. There are no load-bearing self-citations, no fitted parameter renamed as a prediction, and no equation or definition in which a derived quantity is equivalent to its own input. Accordingly, the circularity score is 0.
Assumptions & free parameters
free parameters (3)
- Handler structure assignment =
direct/parallel/serial per tool
- Few-shot exemplars for each prompt =
hand-crafted, not shown in full
- Preference-based hint (TUMS-PRE) =
hand-crafted per dataset
assumptions (3)
- domain assumption The prompted LLM reliably outputs the required structured format so the parser can extract tool and parameters.
- domain assumption ToolQA's tools execute correctly when called with generated parameters.
- domain assumption ToolQA benchmark scores are a meaningful measure of tool-use ability.
Cite this review
Pith. "Pith review of TUMS: Enhancing Tool-use Abilities of LLMs with Multi-structure Handlers." pith.science (2026). https://pith.science/paper/4NVI6EBP
@misc{pith2026250508402,
author = {Pith},
title = {Pith review of: TUMS: Enhancing Tool-use Abilities of LLMs with Multi-structure Handlers},
year = {2026},
howpublished = {\url{https://pith.science/paper/4NVI6EBP}},
note = {Machine review of arXiv:2505.08402}
}
read the original abstract
Recently, large language models(LLMs) have played an increasingly important role in solving a wide range of NLP tasks, leveraging their capabilities of natural language understanding and generating. Integration with external tools further enhances LLMs' effectiveness, providing more precise, timely, and specialized responses. However, LLMs still encounter difficulties with non-executable actions and improper actions, which are primarily attributed to incorrect parameters. The process of generating parameters by LLMs is confined to the tool level, employing the coarse-grained strategy without considering the different difficulties of various tools. To address this issue, we propose TUMS, a novel framework designed to enhance the tool-use capabilities of LLMs by transforming tool-level processing into parameter-level processing. Specifically, our framework consists of four key components: (1) an intent recognizer that identifies the user's intent to help LLMs better understand the task; (2) a task decomposer that breaks down complex tasks into simpler subtasks, each involving a tool call; (3) a subtask processor equipped with multi-structure handlers to generate accurate parameters; and (4) an executor. Our empirical studies have evidenced the effectiveness and efficiency of the TUMS framework with an average of 19.6\% and 50.6\% improvement separately on easy and hard benchmarks of ToolQA, meanwhile, we demonstrated the key contribution of each part with ablation experiments, offering more insights and stimulating future research on Tool-augmented LLMs.
Figures
Figures from the paper (2 more)
Reference graph
Works this paper leans on
-
[1]
arXiv preprint arXiv:2303.08774 (2023)
Achiam, J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F.L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S., et al.: Gpt-4 technical report. arXiv preprint arXiv:2303.08774 (2023)
arXiv 2023
-
[2]
In: Proceed- ings of the AAAI Conference on Artificial Intelligence
Besta, M., Blach, N., Kubicek, A., Gerstenberger, R., Podstawski, M., Giani- nazzi, L., Gajda, J., Lehmann, T., Niewiadomski, H., Nyczyk, P., et al.: Graph of thoughts: Solving elaborate problems with large language models. In: Proceed- ings of the AAAI Conference on Artificial Intelligence. vol. 38, pp. 17682–17690 (2024)
2024
-
[3]
Carlini, N., Ippolito, D., Jagielski, M., Lee, K., Tramer, F., Zhang, C.: Quantifying memorization across neural language models. In: The Eleventh International Conference on Learning Representations (2023), https://openreview.net/forum?id=TatRHT_1cK
work page 2023
-
[4]
arXiv preprint arXiv:2405.05955 (2024)
Chen, J., Liang, J., Wang, B.: Smurfs: Leveraging multiple proficiency agents with context-efficiency for tool planning. arXiv preprint arXiv:2405.05955 (2024)
arXiv 2024
-
[5]
arXiv preprint arXiv:2107.03374 (2021)
Chen, M., Tworek, J., Jun, H., Yuan, Q., Pinto, H.P.d.O., Kaplan, J., Edwards, H., Burda, Y., Joseph, N., Brockman, G., et al.: Evaluating large language models trained on code. arXiv preprint arXiv:2107.03374 (2021)
arXiv 2021
-
[6]
arXiv preprint arXiv:2110.14168 (2021)
Cobbe, K., Kosaraju, V., Bavarian, M., Chen, M., Jun, H., Kaiser, L., Plappert, M., Tworek, J., Hilton, J., Nakano, R., et al.: Training verifiers to solve math word problems. arXiv preprint arXiv:2110.14168 (2021)
arXiv 2021
-
[7]
arXiv preprint arXiv:2301.00234 (2022)
Dong, Q., Li, L., Dai, D., Zheng, C., Wu, Z., Chang, B., Sun, X., Xu, J., Sui, Z.: A survey on in-context learning. arXiv preprint arXiv:2301.00234 (2022)
arXiv 2022
-
[8]
arXiv preprint arXiv:2402.04253 (2024)
Du, Y., Wei, F., Zhang, H.: Anytool: Self-reflective, hierarchical agents for large- scale api calls. arXiv preprint arXiv:2402.04253 (2024)
arXiv 2024
Show all 45 references
-
[9]
arXiv preprint arXiv:2212.04092 (2022)
Dua, D., Gupta, S., Singh, S., Gardner, M.: Successive prompting for decomposing complex questions. arXiv preprint arXiv:2212.04092 (2022)
2022 arXiv
-
[10]
arXiv preprint arXiv:2312.10997 (2023)
Gao, Y., Xiong, Y., Gao, X., Jia, K., Pan, J., Bi, Y., Dai, Y., Sun, J., Wang, H.: Retrieval-augmented generation for large language models: A survey. arXiv preprint arXiv:2312.10997 (2023)
2023 arXiv
-
[11]
In: Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition
Gupta, T., Kembhavi, A.: Visual programming: Compositional visual reasoning without training. In: Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition. pp. 14953–14962 (2023)
2023
-
[12]
Advances in neural information processing systems 36 (2024)
Hao, S., Liu, T., Wang, Z., Hu, Z.: Toolkengpt: Augmenting frozen language models with massive tools via tool embeddings. Advances in neural information processing systems 36 (2024)
2024
-
[13]
arXiv preprint arXiv:2308.00675 (2023)
Hsieh, C.Y., Chen, S.A., Li, C.L., Fujii, Y., Ratner, A., Lee, C.Y., Krishna, R., Pfister, T.: Tool documentation enables zero-shot tool-usage with large language models. arXiv preprint arXiv:2308.00675 (2023)
2023 arXiv
-
[14]
arXiv preprint arXiv:2404.00450 (2024)
Huang, T., Jung, D., Chen, M.: Planning and editing what you retrieve for en- hanced tool learning. arXiv preprint arXiv:2404.00450 (2024)
2024 arXiv
-
[15]
arXiv preprint arXiv:2305.09645 (2023)
Jiang, J., Zhou, K., Dong, Z., Ye, K., Zhao, W.X., Wen, J.R.: Structgpt: A general framework for large language model to reason over structured data. arXiv preprint arXiv:2305.09645 (2023)
2023 arXiv
-
[16]
arXiv preprint arXiv:2210.02406 (2022)
Khot, T., Trivedi, H., Finlayson, M., Fu, Y., Richardson, K., Clark, P., Sabharwal, A.: Decomposed prompting: A modular approach for solving complex tasks. arXiv preprint arXiv:2210.02406 (2022)
2022 arXiv
-
[17]
arXiv preprint arXiv:2311.11315 (2023) 14 A
Kong, Y., Ruan, J., Chen, Y., Zhang, B., Bao, T., Shi, S., Du, G., Hu, X., Mao, H., Li, Z., et al.: Tptu-v2: Boosting task planning and tool usage of large language model-based agents in real-world systems. arXiv preprint arXiv:2311.11315 (2023) 14 A. He et al
2023 arXiv
-
[18]
arXiv preprint arXiv:2403.00839 (2024)
Liu, X., Peng, Z., Yi, X., Xie, X., Xiang, L., Liu, Y., Xu, D.: Toolnet: Con- necting large language models with massive tools via tool graph. arXiv preprint arXiv:2403.00839 (2024)
2024 arXiv
-
[19]
arXiv preprint arXiv:2310.17796 (2023)
Liu, Z., Lai, Z., Gao, Z., Cui, E., Li, Z., Zhu, X., Lu, L., Chen, Q., Qiao, Y., Dai, J., et al.: Controlllm: Augment language models with tools by searching on graphs. arXiv preprint arXiv:2310.17796 (2023)
2023 arXiv
-
[20]
Advances in Neural Information Processing Systems36 (2024)
Lu, P., Peng, B., Cheng, H., Galley, M., Chang, K.W., Wu, Y.N., Zhu, S.C., Gao, J.: Chameleon: Plug-and-play compositional reasoning with large language models. Advances in Neural Information Processing Systems36 (2024)
2024
-
[21]
arXiv preprint arXiv:2402.14158 (2024)
Mekala, D., Weston, J., Lanchantin, J., Raileanu, R., Lomeli, M., Shang, J., Dwivedi-Yu, J.: Toolverifier: Generalization to new tools via self-verification. arXiv preprint arXiv:2402.14158 (2024)
2024 arXiv
-
[22]
Advances in neural information processing sys- tems 35, 27730–27744 (2022)
Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., et al.: Training language models to follow instructions with human feedback. Advances in neural information processing sys- tems 35, 27730–27744 (2022)
2022
-
[23]
arXiv preprint arXiv:2205.12255 (2022)
Parisi, A., Zhao, Y., Fiedel, N.: Talm: Tool augmented language models. arXiv preprint arXiv:2205.12255 (2022)
2022 arXiv
-
[24]
arXiv preprint arXiv:2304.08354 (2023)
Qin, Y., Hu, S., Lin, Y., Chen, W., Ding, N., Cui, G., Zeng, Z., Huang, Y., Xiao, C., Han, C., et al.: Tool learning with foundation models. arXiv preprint arXiv:2304.08354 (2023)
2023 arXiv
-
[25]
arXiv preprint arXiv:2307.16789 (2023)
Qin, Y., Liang, S., Ye, Y., Zhu, K., Yan, L., Lu, Y., Lin, Y., Cong, X., Tang, X., Qian, B., et al.: Toolllm: Facilitating large language models to master 16000+ real-world apis. arXiv preprint arXiv:2307.16789 (2023)
2023 arXiv
-
[26]
arXiv preprint arXiv:2405.16089 (2024)
Qu, C., Dai, S., Wei, X., Cai, H., Wang, S., Yin, D., Xu, J., Wen, J.R.: Colt: Towards completeness-oriented tool retrieval for large language models. arXiv preprint arXiv:2405.16089 (2024)
2024 arXiv
-
[27]
arXiv preprint arXiv:2307.11768 (2023)
Radhakrishnan, A., Nguyen, K., Chen, A., Chen, C., Denison, C., Hernandez, D., Durmus, E., Hubinger, E., Kernion, J., Lukoši¯ ut˙ e, K., et al.: Question decom- position improves the faithfulness of model-generated reasoning. arXiv preprint arXiv:2307.11768 (2023)
2023 arXiv
-
[28]
Advances in Neural Information Processing Systems 36 (2024)
Shen, Y., Song, K., Tan, X., Li, D., Lu, W., Zhuang, Y.: Hugginggpt: Solving ai tasks with chatgpt and its friends in hugging face. Advances in Neural Information Processing Systems 36 (2024)
2024
-
[30]
arXiv preprint arXiv:2403.03031 (2024)
Shi, Z., Gao, S., Chen, X., Yan, L., Shi, H., Yin, D., Chen, Z., Ren, P., Verberne, S., Ren, Z.: Learning to use tools via cooperative and interactive agents. arXiv preprint arXiv:2403.03031 (2024)
2024 arXiv
-
[31]
In: 2023 IEEE International Conference on Robotics and Automation (ICRA)
Singh, I., Blukis, V., Mousavian, A., Goyal, A., Xu, D., Tremblay, J., Fox, D., Thomason, J., Garg, A.: Progprompt: Generating situated robot task plans using large language models. In: 2023 IEEE International Conference on Robotics and Automation (ICRA). pp. 11523–11530. IEEE (2023)
2023
-
[32]
arXiv preprint arXiv:2306.06624 (2023)
Song, Y., Xiong, W., Zhu, D., Li, C., Wang, K., Tian, Y., Li, S.: Restgpt: Con- necting large language models with real-world applications via restful apis. arXiv preprint arXiv:2306.06624 (2023)
2023 arXiv
-
[33]
In: Proceedings of the IEEE/CVF International Conference on Computer Vision
Surís, D., Menon, S., Vondrick, C.: Vipergpt: Visual inference via python execu- tion for reasoning. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 11888–11898 (2023) TUMS: Enhancing Tool-use Abilities of LLMs with Multi-structure Handlers 15
2023
-
[34]
Team, Q.: Introducing qwen1.5 (February 2024), https://qwenlm.github.io/blog/qwen1.5/
2024
-
[35]
arXiv preprint arXiv:2307.09288 (2023)
Touvron, H., Martin, L., Stone, K., Albert, P., Almahairi, A., Babaei, Y., Bash- lykov, N., Batra, S., Bhargava, P., Bhosale, S., et al.: Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288 (2023)
2023 arXiv
-
[36]
arXiv preprint arXiv:2203.11171 (2022)
Wang, X., Wei, J., Schuurmans, D., Le, Q., Chi, E., Narang, S., Chowdhery, A., Zhou, D.: Self-consistency improves chain of thought reasoning in language models. arXiv preprint arXiv:2203.11171 (2022)
2022 arXiv
-
[37]
Advances in neural information processing systems35, 24824–24837 (2022)
Wei, J., Wang, X., Schuurmans, D., Bosma, M., Xia, F., Chi, E., Le, Q.V., Zhou, D., et al.: Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems35, 24824–24837 (2022)
2022
-
[38]
arXiv preprint arXiv:2303.04671 (2023)
Wu, C., Yin, S., Qi, W., Wang, X., Tang, Z., Duan, N.: Visual chatgpt: Talking, drawing and editing with visual foundation models. arXiv preprint arXiv:2303.04671 (2023)
2023 arXiv
-
[39]
arXiv preprint arXiv:2401.00812 (2024)
Yang, K., Liu, J., Wu, J., Yang, C., Fung, Y.R., Li, S., Huang, Z., Cao, X., Wang, X., Wang, Y., et al.: If llm is the wizard, then code is the wand: A survey on how code empowers large language models to serve as intelligent agents. arXiv preprint arXiv:2401.00812 (2024)
2024 arXiv
-
[40]
Advances in Neural Information Processing Systems36 (2024)
Yao, S., Yu, D., Zhao, J., Shafran, I., Griffiths, T., Cao, Y., Narasimhan, K.: Tree of thoughts: Deliberate problem solving with large language models. Advances in Neural Information Processing Systems36 (2024)
2024
-
[41]
arXiv preprint arXiv:2210.03629 (2022)
Yao, S., Zhao, J., Yu, D., Du, N., Shafran, I., Narasimhan, K., Cao, Y.: React: Syn- ergizing reasoning and acting in language models. arXiv preprint arXiv:2210.03629 (2022)
2022 arXiv
-
[42]
arXiv preprint arXiv:2309.17428 (2023)
Yuan, L., Chen, Y., Wang, X., Fung, Y.R., Peng, H., Ji, H.: Craft: Customiz- ing llms by creating and retrieving from specialized toolsets. arXiv preprint arXiv:2309.17428 (2023)
2023 arXiv
-
[43]
arXiv preprint arXiv:2402.19473 (2024)
Zhao, P., Zhang, H., Yu, Q., Wang, Z., Geng, Y., Fu, F., Yang, L., Zhang, W., Cui, B.: Retrieval-augmented generation for ai-generated content: A survey. arXiv preprint arXiv:2402.19473 (2024)
2024 arXiv
-
[44]
arXiv preprint arXiv:2205.10625 (2022)
Zhou, D., Schärli, N., Hou, L., Wei, J., Scales, N., Wang, X., Schuurmans, D., Cui, C., Bousquet, O., Le, Q., et al.: Least-to-most prompting enables complex reasoning in large language models. arXiv preprint arXiv:2205.10625 (2022)
2022 arXiv
-
[45]
arXiv preprint arXiv:2310.13227 (2023)
Zhuang, Y., Chen, X., Yu, T., Mitra, S., Bursztyn, V., Rossi, R.A., Sarkhel, S., Zhang, C.: Toolchain*: Efficient action space navigation in large language models with a* search. arXiv preprint arXiv:2310.13227 (2023)
2023 arXiv
-
[46]
Advances in Neural Information Processing Systems 36 (2024)
Zhuang, Y., Yu, Y., Wang, K., Sun, H., Zhang, C.: Toolqa: A dataset for llm question answering with external tools. Advances in Neural Information Processing Systems 36 (2024)
2024
Reviewed August 15, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.