mlsanityData to 7 Oct 2026

ToolGrad: Efficient Tool-use Dataset Generation with Textual "Gradients"

arXiv 2508.04086 ↗First posted 6 Aug 2025cs.CLcited by 4

On the map: LLM agents and tool use → Tool learning and function calling

Abstract (arXiv)

Prior work synthesizes tool-use LLM datasets by first generating a user query, followed by complex tool-use annotations like depth-first search (DFS). This leads to inevitable annotation failures and low efficiency in data generation. We introduce ToolGrad, an agentic framework that inverts this paradigm. ToolGrad first constructs valid tool-use chains through an iterative process guided by textual "gradients", and then synthesizes corresponding user queries. This "answer-first" approach led to ToolGrad-500, a dataset generated with more complex tool use, lower cost, and almost 100% pass rate. Experiments show that ToolGrad models outperform those trained on expensive baseline datasets and proprietary LLMs. The ToolGrad source code, dataset, and models are available at https://github.com/zhongyi-zhou/toolgrad.

Key facts. Cited by 4 papers in the map (data through 7 Oct 2026). Placed in Tool learning and function calling, one of 16 directions in LLM agents and tool use. Builds on 26 of its 49 references that are in the map. First posted on arXiv 6 Aug 2025.

Builds on

26 of 49 references.

Closest ancestors · 9

Same direction, newest first

  1. 2024Hammer: Robust Function-Calling for On-Device Language Models via Function Masking6 Oct
  2. 2024ToolPlanner: A Tool Augmented LLM for Multi Granularity Instructions with Path Planning and Feedback23 Sep
  3. 2023MetaTool Benchmark for Large Language Models: Deciding Whether to Use Tools and Which to Use4 Oct
  4. 2023ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs31 Jul
  5. 2023ToolQA: A Dataset for LLM Question Answering with External Tools23 Jun
  6. 2023ToolAlpaca: Generalized Tool Learning for Language Models with 3000 Simulated Cases8 Jun
  7. 2023GPT4Tools: Teaching Large Language Model to Use Tools via Self-instruction30 May
  8. 2023Gorilla: Large Language Model Connected with Massive APIs24 May
  9. 2023Toolformer: Language Models Can Teach Themselves to Use Tools9 Feb

Other references · 17

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,160
  2. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,355
  3. 2023Generative Agents: Interactive Simulacra of Human Behaviorcited by 3,500
  4. 2021WebGPT: Browser-assisted question-answering with human feedbackcited by 1,722
  5. 2024τ-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domainscited by 1,280
  6. 2021Retrieval Augmentation Reduces Hallucination in Conversationcited by 691
  7. 2023Automatic Prompt Optimization with "Gradient Descent" and Beam Searchcited by 628
  8. 2023ViperGPT: Visual Inference via Python Execution for Reasoningcited by 591
  9. 2022Visual Programming: Compositional visual reasoning without trainingcited by 507
  10. 2023Describe, Explain, Plan and Select: Interactive Planning with Large Language Models Enables Open-World Multi-Task Agentscited by 396
  11. 2025ToolRL: Reward is All Tool Learning Needscited by 342
  12. 2023Factuality Challenges in the Era of Large Language Modelscited by 125
5 more
  1. 2025Improving LLM-as-a-Judge Inference with the Judgment Distributioncited by 36
  2. 2024MATHSENSEI: A Tool-Augmented Large Language Model for Mathematical Reasoningcited by 34
  3. 2024BlenderLLM: Training Large Language Models for Computer-Aided Design with Self-improvementcited by 26
  4. 2024Adaptive Query Rewriting: Aligning Rewriters through Marginal Probability of Conversational Answerscited by 15
  5. 2025Can Textual Gradient Work in Federated Learning?cited by 14

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 1; 2026: 3.

4 papers.

  1. 2026Execution-First Synthetic Tool-Use Trace Generation for LLM Agentscited by 1
  2. 2025XR Blocks: Accelerating Human-centered AI + XR Innovationcited by 0
  3. 2026Linguistic and Argument Diversity in Synthetic Data for Function-Calling Agentscited by 0
  4. 2026A Matter of TASTE: Improving Coverage and Difficulty of Agent Benchmarkscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024ToolGen: Unified Tool Retrieval and Calling via Generation
  2. 2025ToolScope: Enhancing LLM Agent Tool Use through Tool Merging and Context-Aware Filtering
  3. 2024Efficient and Scalable Estimation of Tool Representations in Vector Space
  4. 2024Toolshed: Scale Tool-Equipped Agents with Advanced RAG-Tool Fusion and Tool Knowledge Bases
  5. 2026Beyond Single-Shot: Multi-step Tool Retrieval via Query Planning
  6. 2025Procedural Environment Generation for Tool-Use Agents
  7. 2025ToolForge: A Data Synthesis Pipeline for Multi-Hop Search without Real-World APIs
  8. 2025ToolMind Technical Report: A Large-Scale, Reasoning-Enhanced Tool-Use Dataset