# Better World Models Can Lead to Better Post-Training Performance

arXiv: https://arxiv.org/abs/2512.03400 · HTML: https://mlsanity.com/paper/2512.03400/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in World models for language agents, one of 11 directions in World models.
- Builds on 18 of its 25 references that are in the map.
- First posted on arXiv 3 Dec 2025.

## Placed in

- World models / [World models for language agents](https://mlsanity.com/wm/world-models-for-language-agents/)
- Mechanistic interpretability / [Circuits and mechanistic analysis](https://mlsanity.com/mechinterp/circuits-and-mechanistic-analysis/)

## Builds on (most cited first, 18 of 18)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [World Models](https://mlsanity.com/paper/1803.10122/) (2018)
- [Understanding intermediate layers using linear classifier probes](https://mlsanity.com/paper/1610.01644/) (2016)
- [Refusal in Language Models Is Mediated by a Single Direction](https://mlsanity.com/paper/2406.11717/) (2024)
- [Recurrent World Models Facilitate Policy Evolution](https://mlsanity.com/paper/1809.01999/) (2018)
- [Discovering Latent Knowledge in Language Models Without Supervision](https://mlsanity.com/paper/2212.03827/) (2022)
- [The Linear Representation Hypothesis and the Geometry of Large Language Models](https://mlsanity.com/paper/2311.03658/) (2023)
- [Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task](https://mlsanity.com/paper/2210.13382/) (2022)
- [Llemma: An Open Language Model For Mathematics](https://mlsanity.com/paper/2310.10631/) (2023)
- [Emergent Linear Representations in World Models of Self-Supervised Sequence Models](https://mlsanity.com/paper/2309.00941/) (2023)
- [Language Models Represent Space and Time](https://mlsanity.com/paper/2310.02207/) (2023)
- [A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity](https://mlsanity.com/paper/2401.01967/) (2024)
- [The Geometry of Categorical and Hierarchical Concepts in Large Language Models](https://mlsanity.com/paper/2406.01506/) (2024)
- [Emergent World Models and Latent Variable Estimation in Chess-Playing Language Models](https://mlsanity.com/paper/2403.15498/) (2024)
- [Shared Global and Local Geometry of Language Model Embeddings](https://mlsanity.com/paper/2503.21073/) (2025)
- [What Does it Mean for a Neural Network to Learn a "World Model"?](https://mlsanity.com/paper/2507.21513/) (2025)

## Built on by (most cited first, 4 of 4)

- [Language Models Struggle to Use Representations Learned In-Context](https://mlsanity.com/paper/2602.04212/) (2026)
- [Toward Consistent World Models with Multi-Token Prediction and Latent Semantic Enhancement](https://mlsanity.com/paper/2604.06155/) (2026)
- [Context-Grounding Gains Are Mediated by Pre-existing Machinery: Auditing GRPO, SFT, and DPO](https://mlsanity.com/paper/2609.00925/) (2026)
- [Does Learning to Predict the World Help Agents Act? Auditing World-Model Post-Training](https://mlsanity.com/paper/2609.33335/) (2026)

## Abstract (arXiv)

We study how explicit world-modeling objectives affect the internal representations and downstream capability of Transformers, using Rubik's Cubes as our training domain. We ask: (1) how does explicitly pretraining a world model affect a model's latent representations, (2) how does world-model quality affect post-training performance, and (3) how should a finite data budget be split between pretraining and task-specific fine-tuning? We compare standard action-prediction fine-tuning with two strategies that add explicit state-prediction supervision: state pretraining followed by fine-tuning, and joint action and state training. We measure task accuracy after Group Relative Policy Optimization (GRPO). We further find that explicit world-modeling yields better representations in terms of higher probing accuracy and steerability of the model, and that better representations yield larger gains from GRPO, especially on harder cube states. Finally, when the fine-tuning budget is held fixed, probe accuracy strongly predicts the GRPO improvement. Under a fixed total data budget, accuracy is maximized by allocating only a small fraction to pretraining.