# IMPLY: Physically Anchored Consistency for World-Model Rollouts

arXiv: https://arxiv.org/abs/2609.12441 · HTML: https://mlsanity.com/paper/2609.12441/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Latent predictive world models, one of 11 directions in World models.
- Builds on 9 of its 11 references that are in the map.
- First posted on arXiv 11 Sep 2026.

## Placed in

- World models / [Latent predictive world models](https://mlsanity.com/wm/latent-predictive-world-models/)

## Builds on (most cited first, 9 of 9)

- [V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning](https://mlsanity.com/paper/2506.09985/) (2025)
- [Verifier-free Test-Time Sampling for Vision-Language-Action Models](https://mlsanity.com/paper/2510.05681/) (2025)
- [Is the Future Compatible? Diagnosing Dynamic Consistency in World Action Models](https://mlsanity.com/paper/2605.07514/) (2026)
- [Hallucination in World Models is Predictable and Preventable](https://mlsanity.com/paper/2606.27326/) (2026)
- [SC3-Eval: Evaluating Robot Foundation Models via Self-Consistent Video Generation](https://mlsanity.com/paper/2606.18610/) (2026)
- [Pre-VLA: Preemptive Runtime Verification for Reliable Vision-Language-Action and World-Model Rollouts](https://mlsanity.com/paper/2605.22446/) (2026)
- [In-Context World Modeling for Robotic Control](https://mlsanity.com/paper/2606.26025/) (2026)
- [Geometry Guided Self-Consistency for Physical AI](https://mlsanity.com/paper/2605.08638/) (2026)
- [CALIPER: Clean Scenes Cannot Rank Physical Inference in Pretrained Visual Representations](https://mlsanity.com/paper/2609.08250/) (2026)

## Abstract (arXiv)

A world model asked what happens if an object is pushed at several speeds produces several futures. If the model has the object in mind, those futures agree about it: each implies the same mass and friction. The consistency checks now used to vet world-action models ask whether a model's futures agree with each other, and none of them knows any physics. We show that this is not enough, and what to do instead. IMPLY reads the physics each rollout implies by inverting a simulator and scores a set of rollouts by how well one object explains all of them, anchored to two calibration pushes the model has observed. In a controlled setting, self-consistency gives a perfect score to a model that ignores the object and always predicts a typical push; anchoring exposes it (AUROC 0.70 versus 1.00). On a real model, V-JEPA 2-AC adapted to the scene, the same thing happens. Given its own calibration pushes the model tracks the object (per-object correlation with the truth 0.91); given another object's, it does not (0.05). Self-consistency cannot tell these apart, preferring the right evidence on 52% of objects, chance level, while anchored disagreement prefers it on 73% and correlates 0.92-0.99 with the rollouts' error. Used to choose among candidate rollout sets, it comes within 0.003 of an oracle that sees the truth. A model that has internalised the wrong object is exactly as self-consistent as one that has internalised the right one; consistency has to be anchored to evidence.