FitText: Evolving Agent Tool Ecologies via Memetic Retrieval
Not on the map. Why?
Abstract (arXiv)
Efficient reasoning is not only a matter of shortening an answer trace; for tool-using agents, it also depends on whether the agent is reasoning over the right action space. As API ecosystems scale to tens of thousands of endpoints, the semantic gap between user requests and tool documentation makes this problem concrete: static retrieval from the initial query can fail before planning begins, and stronger planning alone cannot recover a missing tool. We study this problem as budgeted test-time retrieval and introduce FitText, a training-free framework that makes the tool interface revisable during execution by generating, refining, and evolving natural-language pseudo-tool descriptions as retrieval probes. FitText supports serial refinement, parallel exploration, and Memetic Retrieval, which adds evolutionary selection, local refinement, and tool memory to avoid redundant search. On StableToolBench (16,464 APIs), Memetic FitText reaches an 84.3% pooled pass rate, improving +26.7 points over static retrieval, +22.2 over Single-Pass, +23.2 over Re-Invoke, and +27.5 over Xu-style root refinement. It leads on every evaluated current model, with gains growing alongside model capability, and produces the largest improvements on ambiguous multi-tool tasks where dynamic re-retrieval restores correct candidates after early mistakes. At 40-way concurrency, parallel population execution keeps batched wall-clock at 1.01x Single-Pass despite the added search work.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 54 of its 60 references that are in the map. First posted on arXiv 4 May 2026.
Builds on
54 of 60 references.
References · 54
- 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,160
- 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,355
- 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,601
- 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,072
- 2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 4,064
- 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,317
- 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 3,155
- 2021SimCSE: Simple Contrastive Learning of Sentence Embeddingscited by 2,168
- 2024Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameterscited by 1,928
- 2023ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIscited by 1,880
- 2021WebGPT: Browser-assisted question-answering with human feedbackcited by 1,722
- 2023Gorilla: Large Language Model Connected with Massive APIscited by 1,302
38 more
- 2023HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Facecited by 1,173
- 2023Large Language Models as Optimizerscited by 830
- 2025AlphaEvolve: A coding agent for scientific and algorithmic discoverycited by 763
- 2024Executable Code Actions Elicit Better LLM Agentscited by 537
- 2022Precise Zero-Shot Dense Retrieval without Relevance Labelscited by 508
- 2024Large Language Models: A Surveycited by 506
- 2025GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learningcited by 449
- 2024Understanding the planning of LLM agents: A surveycited by 414
- 2023Tool Learning with Foundation Modelscited by 302
- 2024Tool Learning with Large Language Models: A Surveycited by 221
- 2023MetaTool Benchmark for Large Language Models: Deciding Whether to Use Tools and Which to Usecited by 209
- 2023ART: Automatic multi-step reasoning and tool-use for large language modelscited by 193
- 2023EvoPrompt: Connecting LLMs with Evolutionary Algorithms Yields Powerful Prompt Optimizerscited by 189
- 2025The Danger of Overthinking: Examining the Reasoning-Action Dilemma in Agentic Taskscited by 150
- 2023Query2doc: Query Expansion with Large Language Modelscited by 146
- 2025Dynamic Cheatsheet: Test-Time Learning with Adaptive Memorycited by 146
- 2023RestGPT: Connecting Large Language Models with Real-World RESTful APIscited by 129
- 2024StableToolBench: Towards Stable Large-Scale Benchmarking on Tool Learning of Large Language Modelscited by 129
- 2024AnyTool: Self-Reflective, Hierarchical Agents for Large-Scale API Callscited by 121
- 2022Task-aware Retrieval with Instructionscited by 107
- 2023ToolChain*: Efficient Action Space Navigation in Large Language Models with A* Searchcited by 103
- 2024ToolGen: Unified Tool Retrieval and Calling via Generationcited by 72
- 2025DeepAgent: A General Reasoning Agent with Scalable Toolsetscited by 69
- 2025The Illusion of Diminishing Returns: Measuring Long Horizon Execution in LLMscited by 62
- 2023ControlLLM: Augment Language Models with Tools by Searching on Graphscited by 59
- 2024Re-Invoke: Tool Invocation Rewriting for Zero-Shot Tool Retrievalcited by 50
- 2025Retrieval Models Aren't Tool-Savvy: Benchmarking Tool Retrieval for Large Language Modelscited by 46
- 2025A Survey of Automatic Prompt Engineering: An Optimization Perspectivecited by 41
- 2024ToolRerank: Adaptive and Hierarchy-Aware Reranking for Tool Retrievalcited by 40
- 2024Advancing Tool-Augmented Large Language Models: Integrating Insights from Errors in Inference Treescited by 38
- 2024Tool-Planner: Task Planning with Clusters across Multiple Toolscited by 32
- 2024Towards Completeness-Oriented Tool Retrieval for Large Language Modelscited by 26
- 2024DMQR-RAG: Diverse Multi-Query Rewriting for RAGcited by 26
- 2024Efficient and Scalable Estimation of Tool Representations in Vector Spacecited by 22
- 2026Large Language Model Reasoning Failurescited by 20
- 2024Enhancing Tool Retrieval with Iterative Feedback from Large Language Modelscited by 17
- 2024Corpus-Steered Query Expansion with Large Language Modelscited by 16
- 2024SEE: Strategic Exploration and Exploitation for Cohesive In-Context Prompt Optimizationcited by 12
50 most cited of 54 All 54 (JSON)
Built on by
Not cited yet.