# DART-SD: Diamond-topology Aware Retrieval and Tuning for Self-Distillation of Multi-Turn Tool-Calling Agents

arXiv: https://arxiv.org/abs/2608.18524 · HTML: https://mlsanity.com/paper/2608.18524/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Tool learning and function calling, one of 16 directions in LLM agents and tool use.
- Builds on 39 of its 43 references that are in the map.
- First posted on arXiv 19 Aug 2026.

## Placed in

- LLM agents and tool use / [Tool learning and function calling](https://mlsanity.com/agents/tool-learning-and-function-calling/)
- Knowledge distillation / [On-policy and self-distillation for LLMs](https://mlsanity.com/distill/on-policy-and-self-distillation-for-llms/)

## Builds on (most cited first, 30 of 39)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://mlsanity.com/paper/2507.06261/) (2025)
- [Let's Verify Step by Step](https://mlsanity.com/paper/2305.20050/) (2023)
- [ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs](https://mlsanity.com/paper/2307.16789/) (2023)
- [The Rise and Potential of Large Language Model Based Agents: A Survey](https://mlsanity.com/paper/2309.07864/) (2023)
- [τ-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains](https://mlsanity.com/paper/2406.12045/) (2024)
- [Instruction-Following Evaluation for Large Language Models](https://mlsanity.com/paper/2311.07911/) (2023)
- [GLM-5: from Vibe Coding to Agentic Engineering](https://mlsanity.com/paper/2602.15763/) (2026)
- [Sequence Level Training with Recurrent Neural Networks](https://mlsanity.com/paper/1511.06732/) (2015)
- [Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models](https://mlsanity.com/paper/2601.18734/) (2026)
- [ToolRL: Reward is All Tool Learning Needs](https://mlsanity.com/paper/2504.13958/) (2025)
- [Tool Learning with Foundation Models](https://mlsanity.com/paper/2304.08354/) (2023)
- [ToolHop: A Query-Driven Benchmark for Evaluating Large Language Models in Multi-Hop Tool Use](https://mlsanity.com/paper/2501.02506/) (2025)
- [RoTBench: A Multi-Level Benchmark for Evaluating the Robustness of Large Language Models in Tool Learning](https://mlsanity.com/paper/2401.08326/) (2024)
- [Distilling LLM Agent into Small Models with Retrieval and Code Tools](https://mlsanity.com/paper/2505.17612/) (2025)
- [Beyond Trajectory-Level Attribution: Graph-Based Credit Assignment for Agentic Reinforcement Learning](https://mlsanity.com/paper/2605.26684/) (2026)
- [Sparse but Critical: A Token-Level Analysis of Distributional Shifts in RLVR Fine-Tuning of LLMs](https://mlsanity.com/paper/2603.22446/) (2026)
- [Feedback-Driven Tool-Use Improvements in Large Language Models via Automated Build Environments](https://mlsanity.com/paper/2508.08791/) (2025)
- [Thinking to Recall: How Reasoning Unlocks Parametric Knowledge in LLMs](https://mlsanity.com/paper/2603.09906/) (2026)
- [Multi-Turn On-Policy Distillation with Prefix Replay](https://mlsanity.com/paper/2607.04763/) (2026)
- [Group-Graph Policy Optimization for Long-Horizon Agentic Reinforcement Learning](https://mlsanity.com/paper/2606.22995/) (2026)
- [TL-Training: A Task-Feature-Based Framework for Training Large Language Models in Tool Use](https://mlsanity.com/paper/2412.15495/) (2024)
- [World of Workflows: A Benchmark for Bringing World Models to Enterprise Systems](https://mlsanity.com/paper/2601.22130/) (2026)
- [How Reliable are LLMs as Knowledge Bases? Re-thinking Facutality and Consistency](https://mlsanity.com/paper/2407.13578/) (2024)
- [MatchTIR: Fine-Grained Supervision for Tool-Integrated Reasoning via Bipartite Matching](https://mlsanity.com/paper/2601.10712/) (2026)
- [HINT-SD: Targeted Hindsight Self-Distillation for Long-Horizon Agents](https://mlsanity.com/paper/2605.17873/) (2026)
- [Building Coding Agents via Entropy-Enhanced Multi-Turn Preference Optimization](https://mlsanity.com/paper/2509.12434/) (2025)
- [The Context Gathering Decision Process: A POMDP Framework for Agentic Search](https://mlsanity.com/paper/2605.07042/) (2026)
- [When Models Judge Themselves: Unsupervised Self-Evolution for Multimodal Reasoning](https://mlsanity.com/paper/2603.21289/) (2026)

## Built on by (most cited first, 2 of 2)

- [Graph-Conditioned On-Policy Agent Distillation from Off-the-Shelf Teachers](https://mlsanity.com/paper/2609.37522/) (2026)
- [AdviSD: Learning to Advise Frontier LLMs via Targeted Multi-Turn Self-Distillation](https://mlsanity.com/paper/2609.38142/) (2026)

## Abstract (arXiv)

Equipping Large Language Models (LLMs) with multi-turn tool-calling capabilities is essential for building autonomous agents. However, progress is fundamentally limited by the reliance on full-length trajectory imitation. For tasks involving multiple order-independent sub-goals, the optimal solution space forms a vast combinatorial diamond lattice. Forcing this rich topology into monolithic trajectories causes a severe topological collapse, indiscriminately penalizing valid alternative explorations and severely degrading policy diversity. To address this, we propose DART-SD (Diamond-topology Aware Retrieval and Tuning for Self-Distillation), a novel framework that shifts the paradigm from global forcing to topology-guided localized correction. DART-SD first models the execution process as a converging Interaction-State Transition Graph (ISTG), faithfully capturing the inherent diamond topology of successful and failed exploratory paths. During autonomous rollouts, the framework identifies the Critical Topological Breakpoint (CTB) and retrieves success-supported recovery references. Finally, we introduce a progressive self-distillation paradigm through CTB-guided localized supervision, ensuring that the training loss is calculated exclusively on the generated recovery steps while strictly protecting the valid reasoning prefix from destructive gradient updates. Experiments on complex multi-turn tool-calling benchmarks demonstrate that DART-SD significantly outperforms traditional full-trajectory baselines.