Hugging Face Daily PapersRuiyang Wang, Hao-Lun Hsu, Swarajh MehtaSep 16, 20261 min readpaperadvanced

GAVEL: Graph World Models for Verified and Efficient Long-Horizon LLM Task Planning

Summary

GAVEL augments LLM‑driven robot planners with an explicit graph world model that verifies actions, repairs violations, and reasons over belief distributions, boosting single‑task success from 41 % to 92 % and multi‑task success from 20 % to 93 % on BEHAVIOR‑1K.

A lightweight graph encoding objects, relations, pre‑conditions, and probabilistic beliefs can serve as a fast verifier and repair engine for LLM‑generated plans.
Separating low‑level feasibility checks from high‑level semantic replanning lets the LLM focus on reasoning‑intensive errors only.
Distributional belief reasoning enables dynamic reordering of subtasks, cutting travel distance by ~5 % versus static ordering.
Empirical results on a benchmark of 600 tasks show >50 % absolute improvement in success rates, demonstrating the practical impact of symbolic‑LLM hybrid planning.

Long‑horizon robot planning with LLMs suffers from feasibility violations and poor handling of partial observability. GAVEL shows that a compact, explicit world model can systematically catch and fix these issues, turning LLM planners from brittle generators into reliable agents without heavy retra…

8/10

llm robotics verification planning graph models

Related reading

From Pretraining to Proficiency: Real-World Subtask RL for Long-Horizon Manipulation with Minimal Human Intervention arxiv.org
The paper presents PARTS, a framework that augments a frozen pretrained robot policy with RL‑learned residuals on selected bottleneck subtasks, using local success rewards and minimal human resets. In real‑world bimanual and single‑arm tasks, PARTS more than doubles success rates with only minutes of robot rollouts, outperforming prior fine‑tuning methods.
Hugging Face Daily Papersarxiv.orgSichang Su, Benjamin Yang, Zhiyun Deng4d1 minpaper
The Router Within: Eliciting Native Skill Routing from a Frozen LLM arxiv.org
The paper introduces Gavel, a method that extracts a frozen LLM's internal routing signal via two trained linear maps, eliminating the need to embed skill descriptions in the prompt. Experiments on Qwen3‑32B show up to 13.4‑point improvements on task benchmarks and higher skill‑use accuracy compared to larger retrieval‑based systems.
Hugging Face Daily Papersarxiv.orgRuishuo Chen, Xun Wang, Yu Chen8d1 minpaper
In-Context Robot Learning with VLM Agents arxiv.org
GPT‑Policy is a framework that lets a large vision‑language model (e.g. GPT‑6 Astra) perform in‑context robot learning: a context compiler extracts visual transitions from demos, the VLM proposes tool actions, and a constrained controller verifies and executes them. Real‑robot experiments show that raw video demos improve success rates even without explicit action labels, and that providing align…
Hugging Face Daily Papersarxiv.orgDongzhou Cheng, Taoran Yi, Ye Fang6d1 minpaper
CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Design arxiv.org
CADWorld is a new benchmark suite of 200 long‑horizon mechanical CAD tasks in FreeCAD, covering sketching, part modeling, assembly, CAM, FEM, and more. Agents interact via screenshots and GUI actions; success is checked by executable validation of the saved CAD artifacts. Seven existing agents achieve at most 17.5 % success versus an 87 % expert baseline, highlighting the gap between GUI competen…
Hugging Face Daily Papersarxiv.orgZihan Dong, Yuanzhe Liu, Zhiyuan Ma8d1 minpaper
RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents arxiv.org
The paper presents RecreationWorld, a five‑platform framework that lets hybrid computer‑use agents learn by recreating the behavior of a running reference, and introduces RecreationBench, a 250‑task benchmark with programmatic and visual assertions. Experiments show GPT‑6 Astra reaches 58.1% overall but struggles with deeper programmatic tests, highlighting gaps in current agents.
Hugging Face Daily Papersarxiv.orgShuai Bai, Jiayong Deng, Yikun Fu4d2 minpaper
Learning to solve hard problems in RL for LLMs by never giving up github.io
The post introduces the *Matthew Effect* in RL‑fine‑tuning of LLMs—performance gains concentrate on tasks the model already solves— and proposes *Never Give Up* (NGU), an adaptive sampling scheme that uses a small k for easy prompts and retries hard prompts with a high‑probability “never give up” loop. Experiments on math (AIME, GSM8k), code (Manufactoria), and larger‑scale setups (DeepScaler) sh…
Hacker News front pagegithub.ioMichael Noukhovitch6d11 minHN1179

GAVEL: Graph World Models for Verified and Efficient Long-Horizon LLM Task Planning

Summary

Related reading

From Pretraining to Proficiency: Real-World Subtask RL for Long-Horizon Manipulation with Minimal Human Interventionarxiv.org

The Router Within: Eliciting Native Skill Routing from a Frozen LLMarxiv.org

In-Context Robot Learning with VLM Agentsarxiv.org

CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Designarxiv.org

RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agentsarxiv.org

Learning to solve hard problems in RL for LLMs by never giving upgithub.io

From Pretraining to Proficiency: Real-World Subtask RL for Long-Horizon Manipulation with Minimal Human Intervention arxiv.org

The Router Within: Eliciting Native Skill Routing from a Frozen LLM arxiv.org

In-Context Robot Learning with VLM Agents arxiv.org

CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Design arxiv.org

RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents arxiv.org

Learning to solve hard problems in RL for LLMs by never giving up github.io