Trendoday

V-JEPA-LM: Video-Language Models That Read Dynamics Beyond Static, Text-Aligned Representations

World-model video encoders such as V-JEPA learn how scenes change (direction, speed, order, state) from video alone. To answer questions, they are joined to a language model trained on captions and qu

  • Upvotes13
–Rank now
#7Peak rank
2Days on list
300Score · 7d
Nov 16, 2026First seen
Rank · 7 days

About this paper

Counterfactual supervision over pixel transformations enables language models to read motion dynamics from frozen predictive video encoders, yielding V-JEPA-LM models that outperform open video LLMs across benchmarks.

World-model video encoders such as V-JEPA learn how scenes change (direction, speed, order, state) from video alone. To answer questions, they are joined to a language model trained on captions and question-answer pairs. But each caption or question names only a few attributes of a clip, so the model learns to retrieve those attributes rather than to read the clip, and the dynamics the encoder carries go unused: on the same items, a linear probe reads speed from frozen V-JEPA features at 87% while the VLM answers at chance. We show where dynamics lives in a frozen encoder: in a subspace shared across videos for predictive encoders such as V-JEPA, but not as a shared component of…

world-model video encodersV-JEPAlanguage modellinear probefrozen V-JEPA featuresVLMdynamics subspacepredictive encodersframe-wise text-aligned encodersSigLIP
Upvotes
13
Authors
3
Published
Nov 13, 2026

Summary and abstract excerpt via Hugging Face Papers ↗ · arXiv.

Trending history

V-JEPA-LM: Video-Language Models That Read Dynamics Beyond Static, Text-Aligned Representations was first seen on Hugging Face · Daily papers on Nov 16, 2026. Since then it has appeared in 15 checks on 2 different days. Its best position was #7 of 30, reached on Nov 17, 01:28 UTC. It dropped off the list after Nov 17, 13:28 UTC.

Rank history

Position on Hugging Face · Daily papers at every check (lower is better).

#1#11#20#30 09:2816:2823:2806:2813:28 Nov 16, 09:28 UTC · #20Nov 16, 11:28 UTC · #12Nov 16, 13:28 UTC · #14Nov 16, 15:28 UTC · #9Nov 16, 17:28 UTC · #10Nov 16, 19:28 UTC · #8Nov 16, 21:28 UTC · #8Nov 16, 23:28 UTC · #8Nov 17, 01:28 UTC · #7Nov 17, 03:28 UTC · #7Nov 17, 05:28 UTC · #7Nov 17, 07:28 UTC · #8Nov 17, 09:28 UTC · #7Nov 17, 11:28 UTC · #10Nov 17, 13:28 UTC · #10

Last 14 days

Best rank per day (UTC).

Nov 9today

Metric history

Platform metric as reported by Hugging Face.

091827 09:2816:2823:2806:2813:28 Nov 16, 09:28 UTC · 12Nov 16, 11:28 UTC · 16Nov 16, 13:28 UTC · 18Nov 16, 15:28 UTC · 21Nov 16, 17:28 UTC · 21Nov 16, 19:28 UTC · 21Nov 16, 21:28 UTC · 21Nov 16, 23:28 UTC · 21Nov 17, 01:28 UTC · 22Nov 17, 03:28 UTC · 23Nov 17, 05:28 UTC · 25Nov 17, 07:28 UTC · 26Nov 17, 09:28 UTC · 27Nov 17, 11:28 UTC · 27Nov 17, 13:28 UTC · 27

Questions

Is V-JEPA-LM: Video-Language Models That Read Dynamics Beyond Static, Text-Aligned Representations trending on HF Daily Papers right now?

Not at the moment. It was last on the list on Nov 17, 13:28 UTC, after 2 days of trending.

What is the highest rank V-JEPA-LM: Video-Language Models That Read Dynamics Beyond Static, Text-Aligned Representations reached?

#7 on Hugging Face · Daily papers, on Nov 17, 01:28 UTC. Trendoday checks this list every 120 minutes.

How long has V-JEPA-LM: Video-Language Models That Read Dynamics Beyond Static, Text-Aligned Representations been trending?

It was first seen on Nov 16, 2026 and has been on the list on 2 days (15 checks) since.