pith. sign in

arxiv: 2605.02411 · v2 · pith:KF6EAGED · submitted 2026-05-04 · cs.AI · cs.IR· cs.LG· cs.MA

FitText: Evolving Agent Tool Ecologies via Memetic Retrieval

Kyle Zheng , Han Zhang , Renliang Sun , Chenchen Ye , Wei Wang This is my paper

pith:KF6EAGEDreviewed 2026-07-01 00:23 UTCmodel grok-4.3open to challenge →

classification cs.AI cs.IRcs.LGcs.MA
keywords tool retrievalagent tool usedynamic retrievalmemetic retrievalAPI ecosystemstest-time evolutionsemantic gaptool memory
0
0 comments X

The pith

The retrieval interface, not planning, limits end-to-end agent performance because static search from the first query cannot track how an agent's understanding of needed tools evolves during execution.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper claims that agents' task understanding changes as they work, yet conventional retrieval locks onto the initial query and therefore becomes the main performance bottleneck on large tool sets. FitText moves retrieval inside the reasoning loop so the agent can generate revisable natural-language guesses about the tool it needs, refine those guesses with feedback from each retrieval attempt, and maintain a memory that avoids repeating searches. The evolutionary selection over candidate descriptions produces measurable gains in ranking quality and task completion without any model training. If the claim holds, agent systems can scale to tens of thousands of tools by improving only the retrieval step rather than the planner.

Core claim

FitText treats retrieval as test-time evolution of hypotheses: the agent produces natural-language pseudo-tool descriptions, refines them iteratively using retrieval feedback, and applies selection pressure through a tool memory that discards redundant candidates. On three ToolRet domains this yields NDCG@5 gains of 2.7 to 10.6 points over static query retrieval across base models; on StableToolBench with 16,464 APIs the memetic variant reaches 84.3 percent pooled pass rate, a 26.7-point absolute improvement over static retrieval.

What carries the argument

Memetic Retrieval, the mechanism that places evolutionary selection pressure on a population of revisable natural-language tool descriptions guided by retrieval feedback and a memory of prior attempts.

If this is right

  • Agents can use much larger API collections without changes to their planning logic.
  • Performance improvements appear consistently across different base models when the same reformulation strategies are applied.
  • The training-free design means existing agents can adopt the method immediately.
  • Tool memory reduces redundant searches during the evolutionary process.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Future agent architectures may gain more from investing in dynamic retrieval interfaces than from further planner sophistication.
  • The same evolutionary refinement loop could be tested on long-horizon tasks where requirements also shift over time.
  • Combining memetic retrieval with modest planning improvements might produce additive gains not measured in the current experiments.

Load-bearing premise

Iterative refinement of natural-language tool descriptions using retrieval feedback will converge on better matches without creating new failure modes such as hallucinated requirements or runaway inference cost.

What would settle it

An ablation that disables refinement after the first retrieval attempt and measures whether the reported NDCG and pass-rate gains disappear on the same benchmarks and models.

Figures

Figures reproduced from arXiv: 2605.02411 by Chenchen Ye, Han Zhang, Kyle Zheng, Renliang Sun, Wei Wang.

Figure 1
Figure 1. Figure 1: Comparison between Previous Query-based Retrieval and Dynamic Retrieval. The view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of our proposed dynamic tool retrieval framework. The figure com view at source ↗
Figure 3
Figure 3. Figure 3: Overview of Memetic Retrieval. The framework progresses through five phases: view at source ↗
Figure 4
Figure 4. Figure 4: Effect of query analysis model on retrieval performance. Using GPT-4.1 as the view at source ↗
Figure 5
Figure 5. Figure 5: Effect of refinement turns on retrieval performance. The largest improvement view at source ↗
read the original abstract

A semantic gap separates how users describe tasks from how tools are documented. As API ecosystems scale to tens of thousands of endpoints, static retrieval from the initial query alone cannot bridge this gap: the agent's understanding of what it needs evolves during execution, but its tool set does not. We identify this retrieval interface, not planning, as the binding constraint on end-to-end agent performance, and introduce FitText, a training-free framework that makes retrieval dynamic by embedding it directly in the agent's reasoning loop. FitText treats retrieval as test-time evolution of hypotheses: the agent generates natural-language pseudo-tool descriptions (revisable beliefs about the tool it needs), refines them iteratively using retrieval feedback, and explores diverse alternatives through stochastic generation. Memetic Retrieval adds evolutionary selection pressure over candidate descriptions, guided by a tool memory that avoids redundant search. On ToolRet (three domains), FitText's reformulation strategies improve NDCG@5 by 2.7 to 10.6 points over static query retrieval across all base models; on StableToolBench (16,464 APIs) with GPT-5.4-mini, Memetic reaches an 84.3% pooled pass rate, a 26.7-point absolute gain over static query retrieval.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 1 minor

Summary. The paper claims that the retrieval interface (not planning) is the binding constraint on end-to-end agent performance in large tool ecosystems. It introduces FitText, a training-free method that embeds dynamic retrieval into the agent's loop by generating, iteratively refining, and memetically selecting natural-language pseudo-tool descriptions; this yields NDCG@5 gains of 2.7–10.6 points on ToolRet and an 84.3% pooled pass rate (26.7-point gain) on StableToolBench with GPT-5.4-mini.

Significance. If the empirical gains hold under proper controls, the work would demonstrate that test-time evolutionary refinement of retrieval hypotheses can substantially improve tool selection without model training or planner changes, with potential applicability to scaling agent systems over tens of thousands of APIs.

major comments (3)
  1. [Abstract] Abstract: The central claim that 'retrieval interface, not planning, [is] the binding constraint' rests on one-sided experiments that fix the underlying agent/planner and vary only the retrieval method. No controlled contrast is reported that tests whether comparable gains could be obtained by enhancing planning (e.g., deeper reasoning loops or search) while keeping retrieval static.
  2. [Abstract] Abstract / Experiments: Benchmark deltas are reported without any mention of variance across runs, statistical significance, number of trials, or whether the NDCG@5 and pass-rate improvements survive changes in the underlying retrieval backend or embedding model; this absence prevents assessment of robustness.
  3. [Abstract] Abstract: The weakest assumption—that iterative refinement of pseudo-tool descriptions will converge without introducing new failure modes (hallucinated requirements, excessive cost)—is not addressed by any reported diagnostic or ablation that isolates these risks.
minor comments (1)
  1. [Abstract] The model name 'GPT-5.4-mini' appears in the abstract; confirm whether this is a typographical reference to an existing model or requires clarification.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive feedback. We address each of the major comments below, indicating planned revisions where appropriate.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central claim that 'retrieval interface, not planning, [is] the binding constraint' rests on one-sided experiments that fix the underlying agent/planner and vary only the retrieval method. No controlled contrast is reported that tests whether comparable gains could be obtained by enhancing planning (e.g., deeper reasoning loops or search) while keeping retrieval static.

    Authors: Our experiments deliberately fix the planner to isolate the impact of the proposed retrieval method. The observed gains indicate that retrieval enhancements can substantially improve performance without planner modifications. We recognize that a direct comparison with enhanced planning would provide stronger evidence for the binding constraint claim. Since conducting such experiments would require significant additional work outside the current scope, we will revise the abstract to more precisely state that the results demonstrate the value of dynamic retrieval rather than definitively proving it is the sole binding constraint. revision: partial

  2. Referee: [Abstract] Abstract / Experiments: Benchmark deltas are reported without any mention of variance across runs, statistical significance, number of trials, or whether the NDCG@5 and pass-rate improvements survive changes in the underlying retrieval backend or embedding model; this absence prevents assessment of robustness.

    Authors: We agree that reporting variance, statistical significance, and robustness across backends is important. In the revised manuscript, we will include standard deviations from multiple runs, specify the number of trials, perform significance tests, and add experiments varying the embedding model and retrieval backend to confirm the improvements hold. revision: yes

  3. Referee: [Abstract] Abstract: The weakest assumption—that iterative refinement of pseudo-tool descriptions will converge without introducing new failure modes (hallucinated requirements, excessive cost)—is not addressed by any reported diagnostic or ablation that isolates these risks.

    Authors: The manuscript does not include explicit diagnostics for these failure modes. While our results did not indicate major issues with hallucinated requirements or excessive costs in the tested settings, we will add a new section or appendix with ablations and analysis addressing these risks, including any observed failure cases and mitigation strategies. revision: yes

Circularity Check

0 steps flagged

No circularity; claims rest on empirical benchmark results without self-referential derivation

full rationale

The paper identifies retrieval as the binding constraint via one-sided experiments that vary only the retrieval interface while fixing the agent, then reports NDCG@5 and pass-rate gains on ToolRet and StableToolBench. No equations, fitted parameters, self-citations, or ansatzes are present that would reduce any central claim to its own inputs by construction. The framework is training-free and evaluated against external static-retrieval baselines, making the derivation self-contained against benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the untested premise that retrieval feedback can be used to refine natural-language tool hypotheses without introducing systematic errors, plus the assumption that the reported benchmark deltas generalize beyond the tested models and domains.

axioms (1)
  • domain assumption Retrieval feedback is a reliable signal for refining pseudo-tool descriptions
    Invoked implicitly when the method is described as using retrieval results to iterate on hypotheses.

pith-pipeline@v0.9.1-grok · 5764 in / 1224 out tokens · 16830 ms · 2026-07-01T00:23:09.906880+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

57 extracted references · 57 canonical work pages · 9 internal anchors

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

    work page

  2. [2]

    GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning

    Lakshya A. Agrawal, Shangyin Tan, Dilara Soylu, Noah Ziems, Rishi Khare, Krista Opsahl-Ong, Arnav Singhvi, Herumb Shandilya, Michael J. Ryan, Meng Jiang, Christopher Potts, Koushik Sen, Alexandros G. Dimakis, Ion Stoica, Dan Klein, Matei Zaharia, and Omar Khattab. Gepa: Reflective prompt evolution can outperform reinforcement learning. arXiv preprint arXi...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  3. [3]

    Task-aware retrieval with instructions

    Akari Asai, Timo Schick, Patrick Lewis, Xilun Chen, Gautier Izacard, Sebastian Riedel, Hannaneh Hajishirzi, and Wen-tau Yih. Task-aware retrieval with instructions. arXiv preprint arXiv:2211.09260, 2023

    work page arXiv 2023

  4. [4]

    Advancing tool-augmented large language models: Integrating insights from errors in inference trees

    Sijia Chen, Yibo Wang, Yi-Feng Wu, Qing-Guo Chen, Zhao Xu, Weihua Luo, Kaifu Zhang, and Lijun Zhang. Advancing tool-augmented large language models: Integrating insights from errors in inference trees. arXiv preprint arXiv:2406.07115, 2025

    work page arXiv 2025

  5. [5]

    Re-invoke: Tool invocation rewriting for zero-shot tool retrieval

    Yanfei Chen, Jinsung Yoon, Devendra Sachan, Qingze Wang, Vincent Cohen-Addad, Mohammadhossein Bateni, Chen-Yu Lee, and Tomas Pfister. Re-invoke: Tool invocation rewriting for zero-shot tool retrieval. In Findings of the Association for Computational Linguistics: EMNLP 2024, pp.\ 4705--4726, 2024

    work page 2024

  6. [6]

    The danger of overthinking: Examining the reasoning-action dilemma in agentic tasks

    Alejandro Cuadron, Dacheng Li, Wenjie Ma, Xingyao Wang, Yichuan Wang, Siyuan Zhuang, Shu Liu, Luis Gaspar Schroeder, Tian Xia, Huanzhi Mao, Nicholas Thumiger, Aditya Desai, Ion Stoica, Ana Klimovic, Graham Neubig, and Joseph E. Gonzalez. The danger of overthinking: Examining the reasoning-action dilemma in agentic tasks. arXiv preprint arXiv:2502.08235, 2025

    work page arXiv 2025

  7. [7]

    See: Strategic exploration and exploitation for cohesive in-context prompt optimization

    Wendi Cui, Zhuohang Li, Hao Sun, Damien Lopez, Kamalika Das, Bradley Malin, Sricharan Kumar, and Jiaxin Zhang. See: Strategic exploration and exploitation for cohesive in-context prompt optimization. arXiv preprint arXiv:2402.11347, 2025

    work page arXiv 2025

  8. [8]

    Anytool: Self-reflective, hierarchical agents for large-scale api calls

    Yu Du, Fangyun Wei, and Hongyang Zhang. Anytool: Self-reflective, hierarchical agents for large-scale api calls. arXiv preprint arXiv:2402.04253, 2024

    work page arXiv 2024

  9. [9]

    Mcp-zero: Proactive toolchain construction for llm agents from scratch

    Xiang Fei, Xiawu Zheng, and Hao Feng. Mcp-zero: Proactive toolchain construction for llm agents from scratch. arXiv preprint arXiv:2506.01056, 2025

    work page arXiv 2025

  10. [10]

    In: Annual Meeting of the Association for Computational Linguistics (2023), https://arxiv.org/abs/2212.10496

    Luyu Gao, Xueguang Ma, Jimmy Lin, and Jamie Callan. Precise zero-shot dense retrieval without relevance labels. arXiv preprint arXiv:2212.10496, 2022

    work page arXiv 2022

  11. [11]

    Simcse: Simple contrastive learning of sentence embeddings

    Tianyu Gao, Xingcheng Yao, and Danqi Chen. Simcse: Simple contrastive learning of sentence embeddings. In 2021 Conference on Empirical Methods in Natural Language Processing, EMNLP 2021, pp.\ 6894--6910. Association for Computational Linguistics (ACL), 2021

    work page 2021

  12. [12]

    EvoPrompt: Connecting LLMs with Evolutionary Algorithms Yields Powerful Prompt Optimizers

    Qingyan Guo, Rui Wang, Junliang Guo, Bei Li, Kaitao Song, Xu Tan, Guoqing Liu, Jiang Bian, and Yujiu Yang. Evoprompt: Connecting llms with evolutionary algorithms yields powerful prompt optimizers. arXiv preprint arXiv:2309.08532, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  13. [13]

    Stabletoolbench: Towards stable large-scale benchmarking on tool learning of large language models

    Zhicheng Guo, Sijie Cheng, Hao Wang, Shihao Liang, Yujia Qin, Peng Li, Zhiyuan Liu, Maosong Sun, and Yang Liu. Stabletoolbench: Towards stable large-scale benchmarking on tool learning of large language models. In ACL (Findings), 2024

    work page 2024

  14. [14]

    Understanding the planning of LLM agents: A survey

    Xu Huang, Weiwen Liu, Xiaolong Chen, Xingmei Wang, Hao Wang, Defu Lian, Yasheng Wang, Ruiming Tang, and Enhong Chen. Understanding the planning of llm agents: A survey. arXiv preprint arXiv:2402.02716, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  15. [15]

    arXiv preprint arXiv:2310.03128 , year=

    Yue Huang, Jiawen Shi, Yuan Li, Chenrui Fan, Siyuan Wu, Qihui Zhang, Yixin Liu, Pan Zhou, Yao Wan, Neil Zhenqiang Gong, et al. Metatool benchmark for large language models: Deciding whether to use tools and which to use. arXiv preprint arXiv:2310.03128, 2023

    work page arXiv 2023

  16. [16]

    Bruce Croft

    Victor Lavrenko and W. Bruce Croft. Relevance based language models. Proceedings of the 24th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, pp.\ 120--127, 2001

    work page 2001

  17. [17]

    Corpus-steered query expansion with large language models

    Yibin Lei, Yu Cao, Tianyi Zhou, Tao Shen, and Andrew Yates. Corpus-steered query expansion with large language models. arXiv preprint arXiv:2402.18031, 2024

    work page arXiv 2024

  18. [18]

    u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt \

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich K \"u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt \"a schel, et al. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in neural information processing systems, 33: 0 9459--9474, 2020

    work page 2020

  19. [19]

    Exploring solution divergence and its effect on large language model problem solving

    Hang Li, Kaiqi Yang, Yucheng Chu, Hui Liu, and Jiliang Tang. Exploring solution divergence and its effect on large language model problem solving. arXiv preprint arXiv:2509.22480, 2025 a

    work page arXiv 2025

  20. [20]

    A survey of automatic prompt engineering: An optimization perspective.arXiv preprint arXiv:2502.11560, 2025

    Wenwu Li, Xiangfeng Wang, Wenhao Li, and Bo Jin. A survey of automatic prompt engineering: An optimization perspective. arXiv preprint arXiv:2502.11560, 2025 b

    work page arXiv 2025

  21. [21]

    arXiv preprint arXiv:2510.21618 , year=

    Xiaoxi Li, Wenxiang Jiao, Jiarui Jin, Guanting Dong, Jiajie Jin, Yinuo Wang, Hao Wang, Yutao Zhu, Ji-Rong Wen, Yuan Lu, and Zhicheng Dou. Deepagent: A general reasoning agent with scalable toolsets. arXiv preprint arXiv:2510.21618, 2025 c

    work page arXiv 2025

  22. [22]

    arXiv preprint arXiv:2411.13154 , year=

    Zhicong Li, Jiahao Wang, Zhishu Jiang, Hangyu Mao, Zhongxia Chen, Jiazhen Du, Yuanxing Zhang, Fuzheng Zhang, Di Zhang, and Yong Liu. Dmqr-rag: Diverse multi-query rewriting for rag. arXiv preprint arXiv:2411.13154, 2024

    work page arXiv 2024

  23. [23]

    Query rewriting via large language models

    Jie Liu and Barzan Mozafari. Query rewriting via large language models. arXiv preprint arXiv:2403.09060, 2024

    work page arXiv 2024

  24. [24]

    Liu, Y ., Wang, W., Feng, R., Zhang, Y ., Xu, G., Deng, G., Li, Y ., and Zhang, L

    Yanming Liu, Xinyue Peng, Jiannan Cao, Shi Bo, Yuwei Zhang, Xuhong Zhang, Sheng Cheng, Xun Wang, Jianwei Yin, and Tianyu Du. Tool-planner: Task planning with clusters across multiple tools. arXiv preprint arXiv:2406.03807, 2025

    work page arXiv 2025

  25. [25]

    Self-Refine: Iterative Refinement with Self-Feedback

    Aman Madaan, Niket Tandon, Prakhar Gupta, Skyler Hallinan, Luyu Gao, Sarah Wiegreffe, Uri Alon, Nouha Dziri, Shrimai Prabhumoye, Yiming Yang, Shashank Gupta, Bodhisattwa Prasad Majumder, Katherine Hermann, Sean Welleck, Amir Yazdanbakhsh, and Peter Clark. Self-refine: Iterative refinement with self-feedback. arXiv preprint arXiv:2303.17651, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  26. [26]

    Large Language Models: A Survey

    Shervin Minaee, Tomas Mikolov, Narjes Nikzad, Meysam Chenaghlu, Richard Socher, Xavier Amatriain, and Jianfeng Gao. Large language models: A survey. arXiv preprint arXiv:2402.06196, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  27. [27]

    Efficient and scalable estimation of tool representations in vector space

    Suhong Moon, Siddharth Jha, Lutfi Eren Erdogan, Sehoon Kim, Woosang Lim, Kurt Keutzer, and Amir Gholami. Efficient and scalable estimation of tool representations in vector space. arXiv preprint arXiv:2409.02141, 2024

    work page arXiv 2024

  28. [28]

    WebGPT: Browser-assisted question-answering with human feedback

    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

    work page internal anchor Pith review Pith/arXiv arXiv 2021

  29. [29]

    Alexander Novikov, Ng \^a n V \ u , Marvin Eisenberger, Emilien Dupont, Po-Sen Huang, Adam Zsolt Wagner, Sergey Shirobokov, Borislav Kozlovskii, Francisco J. R. Ruiz, Abbas Mehrabian, M. Pawan Kumar, Abigail See, Swarat Chaudhuri, George Holland, Alex Davies, Sebastian Nowozin, Pushmeet Kohli, and Matej Balog. Alphaevolve: A coding agent for scientific an...

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  30. [30]

    ART: Automatic multi-step reasoning and tool-use for large language models

    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

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  31. [31]

    Gorilla: Large language model connected with massive apis

    Shishir G Patil, Tianjun Zhang, Xin Wang, and Joseph E Gonzalez. Gorilla: Large language model connected with massive apis. Advances in Neural Information Processing Systems, 37: 0 126544--126565, 2024

    work page 2024

  32. [32]

    Tool learning with foundation models

    Yujia Qin, Shengding Hu, Yankai Lin, Weize Chen, Ning Ding, Ganqu Cui, Zheni Zeng, Xuanhe Zhou, Yufei Huang, Chaojun Xiao, et al. Tool learning with foundation models. ACM Computing Surveys, 57 0 (4): 0 1--40, 2024 a

    work page 2024

  33. [33]

    Toolllm: Facilitating large language models to master 16000+ real-world apis

    Yujia Qin, Shihao Liang, Yining Ye, Kunlun Zhu, Lan Yan, Yaxi Lu, Yankai Lin, Xin Cong, Xiangru Tang, Bill Qian, et al. Toolllm: Facilitating large language models to master 16000+ real-world apis. In The Twelfth International Conference on Learning Representations, 2024 b

    work page 2024

  34. [34]

    Colt: Towards completeness-oriented tool retrieval for large language models

    Changle Qu, Sunhao Dai, Xiaochi Wei, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, Jun Xu, and Ji-Rong Wen. Colt: Towards completeness-oriented tool retrieval for large language models. arXiv e-prints, pp.\ arXiv--2405, 2024 a

    work page 2024

  35. [35]

    Towards completeness-oriented tool retrieval for large language models

    Changle Qu, Sunhao Dai, Xiaochi Wei, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, Jun Xu, and Ji-Rong Wen. Towards completeness-oriented tool retrieval for large language models. In Proceedings of the 33rd ACM International Conference on Information and Knowledge Management, pp.\ 1930--1940, 2024 b

    work page 1930

  36. [36]

    Tool learning with large language models: A survey

    Changle Qu, Sunhao Dai, Xiaochi Wei, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, Jun Xu, and Ji-Rong Wen. Tool learning with large language models: A survey. Frontiers of Computer Science, 19 0 (8): 0 198343, 2025

    work page 2025

  37. [37]

    The probabilistic relevance framework: Bm25 and beyond

    Stephen Robertson, Hugo Zaragoza, et al. The probabilistic relevance framework: Bm25 and beyond. Foundations and Trends in Information Retrieval , 3 0 (4): 0 333--389, 2009

    work page 2009

  38. [38]

    Toolformer: Language models can teach themselves to use tools

    Timo Schick, Jane Dwivedi-Yu, Roberto Dess \` , Roberta Raileanu, Maria Lomeli, Eric Hambro, Luke Zettlemoyer, Nicola Cancedda, and Thomas Scialom. Toolformer: Language models can teach themselves to use tools. Advances in Neural Information Processing Systems, 36: 0 68539--68551, 2023

    work page 2023

  39. [39]

    Tooldreamer: Instilling llm reasoning into tool retrievers,

    Saptarshi Sengupta, Zhengyu Zhou, Jun Araki, Xingbo Wang, Bingqing Wang, Suhang Wang, and Zhe Feng. Tooldreamer: Instilling llm reasoning into tool retrievers. arXiv preprint arXiv:2510.19791, 2025

    work page arXiv 2025

  40. [40]

    Hugginggpt: Solving ai tasks with chatgpt and its friends in hugging face

    Yongliang Shen, Kaitao Song, Xu Tan, Dongsheng Li, Weiming Lu, and Yueting Zhuang. Hugginggpt: Solving ai tasks with chatgpt and its friends in hugging face. Advances in Neural Information Processing Systems, 36: 0 38154--38180, 2023

    work page 2023

  41. [41]

    Retrieval models aren’t tool-savvy: Benchmarking tool retrieval for large language models,

    Zhengliang Shi, Yuhan Wang, Lingyong Yan, Pengjie Ren, Shuaiqiang Wang, Dawei Yin, and Zhaochun Ren. Retrieval models aren't tool-savvy: Benchmarking tool retrieval for large language models. arXiv preprint arXiv:2503.01763, 2025

    work page arXiv 2025

  42. [42]

    Sinha et al

    Akshit Sinha, Arvindh Arun, Shashwat Goel, Steffen Staab, and Jonas Geiping. The illusion of diminishing returns: Measuring long horizon execution in llms. arXiv preprint arXiv:2509.09677, 2026

    work page arXiv 2026

  43. [43]

    Large language model reasoning failures

    Peiyang Song, Pengrui Han, and Noah Goodman. Large language model reasoning failures. arXiv preprint, 2026

    work page 2026

  44. [44]

    Restgpt: Connecting large language models with real-world restful apis.arXiv preprint arXiv:2306.06624,

    Yifan Song, Weimin Xiong, Dawei Zhu, Wenhao Wu, Han Qian, Mingbo Song, Hailiang Huang, Cheng Li, Ke Wang, Rong Yao, et al. Restgpt: Connecting large language models with real-world restful apis. arXiv preprint arXiv:2306.06624, 2023

    work page arXiv 2023

  45. [45]

    A statistical interpretation of term specificity and its application in retrieval

    Karen Sparck Jones. A statistical interpretation of term specificity and its application in retrieval. Journal of documentation, 28 0 (1): 0 11--21, 1972

    work page 1972

  46. [46]

    Dynamic Cheatsheet: Test-Time Learning with Adaptive Memory

    Mirac Suzgun, Mert Yuksekgonul, Federico Bianchi, Dan Jurafsky, and James Zou. Dynamic cheatsheet: Test-time learning with adaptive memory. arXiv preprint arXiv:2504.07952, 2025

    work page Pith review arXiv 2025

  47. [47]

    Query2doc: Query expansion with large language models

    Liang Wang, Nan Yang, and Furu Wei. Query2doc: Query expansion with large language models. arXiv preprint arXiv:2303.07678, 2023

    work page arXiv 2023

  48. [48]

    Toolgen: Unified tool retrieval and calling via generation

    Renxi Wang, Xudong Han, Lei Ji, Shu Wang, Timothy Baldwin, and Haonan Li. Toolgen: Unified tool retrieval and calling via generation. arXiv preprint arXiv:2410.03439, 2025

    work page arXiv 2025

  49. [49]

    Executable code actions elicit better llm agents

    Xingyao Wang, Yangyi Chen, Lifan Yuan, Yizhe Zhang, Yunzhu Li, Hao Peng, and Heng Ji. Executable code actions elicit better llm agents. In Forty-first International Conference on Machine Learning, 2024

    work page 2024

  50. [50]

    Enhancing tool retrieval with iterative feedback from large language models

    Qiancheng Xu, Yongqi Li, Heming Xia, and Wenjie Li. Enhancing tool retrieval with iterative feedback from large language models. In Findings of the Association for Computational Linguistics: EMNLP 2024, pp.\ 9609--9619, 2024

    work page 2024

  51. [51]

    Large Language Models as Optimizers

    Chengrun Yang, Xuezhi Wang, Yifeng Lu, Hanxiao Liu, Quoc V. Le, Denny Zhou, and Xinyun Chen. Large language models as optimizers. arXiv preprint arXiv:2309.03409, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  52. [52]

    React: Synergizing reasoning and acting in language models

    Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik R Narasimhan, and Yuan Cao. React: Synergizing reasoning and acting in language models. In The eleventh international conference on learning representations, 2022

    work page 2022

  53. [53]

    Toolrerank: Adaptive and hierarchy-aware reranking for tool retrieval

    Yuanhang Zheng, Peng Li, Wei Liu, Yang Liu, Jian Luan, and Bin Wang. Toolrerank: Adaptive and hierarchy-aware reranking for tool retrieval. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), pp.\ 16263--16273, 2024

    work page 2024

  54. [54]

    arXiv preprint arXiv:2310.13227 , year=

    Yuchen Zhuang, Xiang Chen, Tong Yu, Saayan Mitra, Victor Bursztyn, Ryan A. Rossi, Somdeb Sarkhel, and Chao Zhang. Toolchain*: Efficient action space navigation in large language models with a* search. arXiv preprint arXiv:2310.13227, 2023

    work page arXiv 2023

  55. [55]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should not add it explicitly Type <Return> for now, but then later remove the command n...

    work page

  56. [56]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@first@sw \@firstoftwo \@ifundefined NAT@b*@#2 \@firstoftwo @num @NAT@ctr \@secondoft...

    work page

  57. [57]

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibsetup #1 @NAT@ctr @ @openbib .11em \@plus.33em \@minus.07em 4000 4000 `\.\@m @bibit...

    work page