From Zero to Hero: Training-Free Custom Concept Spawning in World Models

arXiv 2026 59.4 method, application

TLDR

Training-free method SPAWN enables custom concept spawning in autoregressive world models by swapping anchor frames.

Reasoning

The paper introduces a novel training-free approach for controllable scene composition, addressing a key limitation of world models. However, the abstract lacks empirical validation or real-world experiments, and the method's generality is unclear.

Read-first score

Read-first score 59.4, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 55.

Recency 6%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Citation impact 18%

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. citation_normalized_percentile=0.91962043

Topical relevance 29%

78.6

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Methodology quality 18%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=experiment

Reproducibility 18%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Citation velocity 12%

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=0.00

Field roles

FoundationFrontierBridge

Rank sensitivity

Stability: volatile; rank range: 501.

Keyword Scores

world model

interactive world model

generative world model

video world model

world simulator

world dynamics prediction

model-based reinforcement learning world model

Deep Analysis

Innovations

Training-free concept spawning in autoregressive world models
Exploiting the pinned anchor in context memory for concept injection
Swapping anchor with external concept latent over a short injection window

Methodology

SPAWN leverages the structural property of image-to-video backbones where the first slot of context memory is pinned to the reference frame. It swaps this anchor with an external concept latent over a short injection window, then returns the original anchor, allowing the concept to propagate through the rollout via the model's own memory. The method accepts either a concept image or a text description as input.

Key Results

SPAWN integrates concepts with consistent lighting, scale, and perspective while preserving identity and temporal coherence, demonstrating that controllable concept spawning is achievable in existing autoregressive world models without any training.

Resources

Paper