Skip to content
AI News HubLIVE
Original source2 min read

Grounding Spatial Relations in a Compact World Model: Instruction Leakage and a Goal-Free Dynamics Fix

Summary

A new arXiv paper shows that goal-conditioned world models can achieve high relation-readout accuracy by transcribing instructions rather than perceiving scenes, and proposes a goal-free dynamics plus supervised read path as the remedy.

SourcearXiv AIAuthor: Yufeng Wang, Lu Wei, Haibin Ling
Grounding Spatial Relations in a Compact World Model: Instruction Leakage and a Goal-Free Dynamics Fix
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 8 Jul 2026]

Title:Grounding Spatial Relations in a Compact World Model: Instruction Leakage and a Goal-Free Dynamics Fix

View a PDF of the paper titled Grounding Spatial Relations in a Compact World Model: Instruction Leakage and a Goal-Free Dynamics Fix, by Yufeng Wang and 2 other authors

View PDF HTML (experimental)

Abstract:Compact world models that condition on a language goal promise to ground relations such as ``put the red block left of the blue block'' using a sparse set of explicit \emph{reference anchors}. We ask when such references actually ground a relation, and identify a trap: a goal-conditioned predictor reaches a striking $0.90$ relation-readout accuracy, yet this is \emph{instruction transcription}, not perception. Withholding the goal collapses it to chance ($0.90\!\to\!0.27$, three seeds) and a counterfactual instruction makes the predicted anchors follow the \emph{false} instruction $94.5\%$ of the time (true scene $2.3\%$; $N{=}256$). Tested across three settings and a within-task ablation, our central claim characterizes the confound: \textbf{instruction leakage occurs when the scored quantity is transcribable from the instruction (when the instruction names the answer) and is essentially independent of how predictive the non-instruction inputs are.} Our tabletop and the external BabyAI benchmark leak, whereas a Language-Table forward-dynamics world model whose instruction names \emph{referents} does not, until the instruction is augmented to name the direction; and degrading the action never increases leakage, the opposite of what predictor-competition predicts. The diagnosis prescribes the fix: keep the goal out of the dynamics (it belongs to the planner's cost) and supervise the \emph{read} path, recovering genuine, instruction-independent grounding ($0.88$, identical with and without the goal). The detection protocol and remedy apply to any goal-conditioned world model whose instruction names the scored quantity.

Subjects:

Artificial Intelligence (cs.AI)

Cite as: arXiv:2607.06925 [cs.AI]

(or arXiv:2607.06925v1 [cs.AI] for this version)

https://doi.org/10.48550/arXiv.2607.06925

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Yufeng Wang [view email] [v1] Wed, 8 Jul 2026 02:38:43 UTC (76 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Grounding Spatial Relations in a Compact World Model: Instruction Leakage and a Goal-Free Dynamics Fix, by Yufeng Wang and 2 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.AI

new | recent | 2026-07

Change to browse by:

cs

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

ResearchersAdvanced

Key points

  • A goal-conditioned predictor hit 0.90 relation accuracy but collapsed to 0.27 when the goal was withheld
  • Counterfactual instructions made predicted anchors follow the false instruction 94.5% of the time versus 2.3% for the true scene
  • Leakage depends on whether the scored quantity is transcribable from the instruction, not on how predictive other inputs are
  • Keeping the goal out of the dynamics and supervising the read path restored instruction-independent grounding at 0.88

Highlights and analysis are generated automatically and may contain errors. Check the original source.