Skip to content
AI News HubLIVE
Original source2 min read

Auditing Learned Dynamics in Weather Models

Summary

A study using sparse autoencoder features from GraphCast finds that causal relevance does not require conventional physical semantics: ablating a mesh-locked feature increases global geopotential height error, while amplifying convection-related features cuts tropical-cyclone intensity error by a median of 51% across seven storms.

SourcearXiv Machine LearningAuthor: Skylar Gu, Christina Last
Auditing Learned Dynamics in Weather Models
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 18 Nov 2026]

Title:Auditing Learned Dynamics in Weather Models

View a PDF of the paper titled Auditing Learned Dynamics in Weather Models, by Skylar Gu and 1 other authors

View PDF HTML (experimental)

Abstract:Learned weather models are experimentally accessible computational systems whose internal representations can be inspected and manipulated. Using sparse autoencoder features from GraphCast, DeepMind's autoregressive weather forecasting model, we test whether causal relevance requires physical semantics, a question that matters for auditing whether AI forecasters used in high-assurance settings rely on reliable mechanisms. We also look at whether physically associated latent directions can be steered to improve extreme-event forecasts, where interventions are applied to the model's internal activations during the rollout. We find that ablating a feature locked to GraphCast's computational mesh increases global geopotential height error despite lacking conventional physical semantics. Second, suppressing convection- and low-level-vorticity-associated features weakens predicted tropical-cyclone deepening (the fall in central pressure during intensification), while amplifying them reduces intensity error, with amplification of convection features through feature scaling achieving a median of 51\% error reduction over seven storms. Together, these results separate interpretable, physical association from causal utility and establish sparse features as experimental coordinates for auditing and intervening on learned weather dynamics.

Subjects:

Machine Learning (cs.LG)

Cite as: arXiv:2611.25119 [cs.LG]

(or arXiv:2611.25119v1 [cs.LG] for this version)

https://doi.org/10.48550/arXiv.2611.25119

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Skylar Gu [view email] [v1] Wed, 18 Nov 2026 19:58:01 UTC (2,961 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Auditing Learned Dynamics in Weather Models, by Skylar Gu and 1 other authors

View PDF

HTML (experimental)

TeX Source

view license

Additional Features

Audio Summary

Current browse context:

cs.LG

new | recent | 2026-11

Change to browse by:

cs

References & Citations

SciX

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?)

IArxiv recommender toggle

IArxiv Recommender (What is IArxiv?)

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

  • The study tests whether causal relevance in GraphCast requires physical semantics using sparse autoencoder features
  • Ablating a feature tied to GraphCast's computational mesh increases global geopotential height error despite lacking conventional physical meaning
  • Suppressing convection and low-level-vorticity features weakens cyclone deepening, while amplifying convection features reduces intensity error by a median of 51% over seven storms
  • The work separates interpretable physical association from causal utility and frames sparse features as experimental coordinates for auditing and intervening on learned weather dynamics

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