Skip to content
AI News HubLIVE
Original source2 min read

The Verifier is the Curriculum: Execution-Gated Self-Distillation for Cross-Family Game Generation

Summary

The paper replaces a learned judge with a deterministic, judge-free strict-launch filter and uses rejection-sampling self-distillation to raise a 14B model's clean generation on four unseen game families from 8.8% to 42.2%, reaching 25/25 best-of-K coverage. A control that swaps in a lenient BUILD check erases the gain, isolating verifier precision as the curriculum signal.

SourcearXiv AIAuthor: Chenyu Zhou, Qiliang Jiang, Shuning Wu, Xu Zhou
The Verifier is the Curriculum: Execution-Gated Self-Distillation for Cross-Family Game Generation
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 23 Jun 2026]

Title:The Verifier is the Curriculum: Execution-Gated Self-Distillation for Cross-Family Game Generation

View a PDF of the paper titled The Verifier is the Curriculum: Execution-Gated Self-Distillation for Cross-Family Game Generation, by Chenyu Zhou and 3 other authors

View PDF HTML (experimental)

Abstract:Post-training a code generator against a learned judge can optimize proxy features that raise the score without improving the artifact. We study the opposite signal: a deterministic, judge-free, ungameable filter -- whether a generated project launches cleanly under a headless engine (strict-launch). Under this gate, rejection-sampling self-distillation compounds out-of-family generalization. On GameCraft-Bench (mapping a natural-language brief to a complete Godot project), a 14B model (Qwen3-14B+LoRA) distilled under strict-launch raises clean generation on four unseen game families from 8.8% to 42.2% per-candidate and best-of-K coverage from 18/25 to 25/25 (the gold ceiling) over three rounds, each a significant gain (p=0.0019, p

new | recent | 2026-07

Change to browse by:

cs cs.SE

References & Citations

NASA ADS

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?)

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

EngineersAdvanced

Key points

  • Three rounds of strict-launch-gated self-distillation on GameCraft-Bench lift clean generation from 8.8% to 42.2% and best-of-K coverage from 18/25 to the 25/25 gold ceiling.
  • An exactly matched gold-duplication control regresses to 5.6%, ruling out added data volume as the explanation.
  • Swapping only the filter to a lenient BUILD check with a 99.9% pass rate removes the gain entirely, showing the effect comes from verifier precision rather than the optimizer.

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