The paper introduces Tool Primitives, a design that replaces rigid API schemas with natural language interfaces for tool calling, enabling seamless inter-tool communication. It builds ToolFace, a repository of over 25,000 functions that LLMs can dynamically retrieve, and HEART, a harness engineering framework that orchestrates tool use with planning, routing, and verification. Experiments show HEART outperforms fine‑tuned models and leading commercial LLMs while cutting API costs by up to 85%.
By Haibo Jin, Suijin Wang, Xucheng Yu, Haojing Luo, Haohan Wang
arXiv:2610.02376v1 Announce Type: cross
Abstract: Co-designing ML models and the accelerators that run them is an unusual reasoning task: architects must draw confident, high-stakes conclusions about...
By Samuel Kushnir, Kavya Sreedhar, Yeshwanth Reddy Pogula, Amir Yazdanbakhsh, Narges Shahidi, Ming Liu, Varun Gohil, Ravi Iyer, Parthasarathy Ranganathan, Christina Delimitrou, Suvinay Subramanian
arXiv:2606. 05268v1 Announce Type: cross Abstract: We present a pipeline for building and aggregating task-specific, LLM-generated weak (imperfect) verifiers into a strong verifier for spatial layout domains.
By Sharon Zhang, R. Kenny Jones, Jiajun Wu, Maneesh Agrawala
arXiv:2608. 02650v1 Announce Type: new Abstract: Large language model (LLM) agents increasingly rely on external tools to complete complex real-world tasks.
By Zian Zhai, Xingyu Tan, Gaowang Zou, Xiaoyang Wang, Wenjie Zhang
The paper introduces domain-specific languages (DSLs) as an explicit representation layer to bridge the gap between high-level natural-language abstractions and concrete actions. Experiments on the visually-grounded “Hexagons” task and the “Text2CAD” benchmark show that code generation using a human-curated DSL yields dramatic performance gains, especially at higher abstraction levels, while automatically induced DSLs only partially recover these gains. The study highlights that the right representation can close much of the performance gap caused by abstraction, but that automatically learning such representations remains a challenge.
By Amit Lisha, Reut Tsarfaty
arXiv:2609.07434v1 Announce Type: new
Abstract: Natural-language Computer-Aided Design (CAD) code generation aims to turn design intent into executable and editable parametric programs. Large languag...
By Yali Du, Hui Sun, San-Zhuo Xi, Ming Li
arXiv:2610.19238v1 Announce Type: new
Abstract: When writing code, the same computation can be expressed in different programming languages and through different implementation strategies. We show th...
By Casper L. Christensen, Kyle J. Ray, Dmitry Manning-Coe, Adam Shai, Paul Riechers
When writing code, the same computation can be expressed in different programming languages and through different implementation strategies. We show that the activations of LLMs capture this shared st...
arXiv:2606. 15943v1 Announce Type: cross Abstract: Engineering LLM-native software remains a challenging and immature field.
By V\'ictor A. Braberman, Flavia Bonomo-Braberman
EvoDesign is an agentic framework that enables editable diagram creation by constructing a diagram manifest through coordinated structure, style, and layout agents, then rendering it in a diagnostic verification environment that provides defect signals and VLM-based critique. The system uses an evolution mechanism to distill refinement traces into a three-tier design expertise memory, promoting strategies only with repeated cross-context support. CanvasBench, a new benchmark and evaluation protocol, assesses content, visual, and cognitive dimensions, and experiments show EvoDesign achieves a strong balance across content fidelity, visual organization, and cognitive usability compared to pixel, code-based, agentic, one-shot schema, and canvas-native methods.
By Tianfu Wang, Leilei Ding, Ziyang Tao, Yi Zhan, Zhiyuan Ma, Wei Wu, Yuxuan Lei, Junyang Wang, Yizhao Xu, Yin WU, Zhengyu Hu, Hongyuan Zhu, Nicholas Jing Yuan, Yanyong Zhang, Hui Xiong
arXiv:2502. 08884v3 Announce Type: replace-cross Abstract: We present ShapeLib, the first method that uses the priors of Large Language Models (LLMs) to design libraries of programmatic 3D shape abstractions.
By R. Kenny Jones, Paul Guerrero, Niloy J. Mitra, Daniel Ritchie
ConceptARC at Scale introduces a generator that produces thousands of ARC-style tasks focused on specific spatial and semantic concepts, expanding the original benchmark by tenfold. The new dataset, ConceptARC-100, contains 1,600 manually reviewed tasks across 16 concepts, each with 100 tasks. Evaluation of large reasoning models on this dataset reveals that they reliably solve certain concepts, struggle with others, and develop internal representations that cluster by transformation operations such as recoloring, removal, or translation.
By Philipp Mondorf, Adrien Deli\`ege, Solim LeGris, Claas Beger, Ryan Yi, Barbara Plank, Brenden M. Lake, Todd M. Gureckis