Metraipolis › archive › dc_65be00a41585

GlyphBench: A Playground for Language-Model Reinforcement Learning — arXiv:2609.34214

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.34214
License (as archived): CC0-1.0
Archived: 2026-10-07T20:57:34.188Z
SHA-256 of the archived text: 0f64c5273a1c823cf5e0a03ceb7167c89f81160faf05eb8ce2688472d199db16
Stable URL of this page: https://metraipolis.com/documento/dc_65be00a41585

How to cite

GlyphBench: A Playground for Language-Model Reinforcement Learning — arXiv:2609.34214 — arXiv (en), https://arxiv.org/abs/2609.34214 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_65be00a41585 (Metraipolis archive dc_65be00a41585, sha256 0f64c5273a1c823cf5e0a03ceb7167c89f81160faf05eb8ce2688472d199db16)

BibTeX

@misc{dc65be00a41585,
  title = {GlyphBench: A Playground for Language-Model Reinforcement Learning — arXiv:2609.34214},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.34214},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_65be00a41585; sha256 0f64c5273a1c823cf5e0a03ceb7167c89f81160faf05eb8ce2688472d199db16},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34214. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We introduce GlyphBench, an environment suite for reinforcement learning (RL) post-training of language-model agents, with over 360 tasks spanning diverse games. GlyphBench renders spatial observations as two-dimensional Unicode grids and connects training, evaluation, and trajectory replay through a unified interface designed to support efficient and reproducible research. We use GlyphBench to study how observation interfaces, reasoning effort, and agent harnesses affect performance, and how RL configurations shape learning dynamics. Our results show that glyph observations outperform native text and pixels in our Craftax experiments, with further gains on several BALROG environments. RL on 100 GlyphBench tasks improves Qwen3.5-4B on held-out Reasoning Gym problems, reaching 63.48% accuracy and outperforming the base model, a math-trained baseline, and a code-trained baseline. These experiments provide empirical evidence that reasoning gains from gameplay can yield stronger transfer than math or code. Together, these results highlight GlyphBench's value as a testbed for systematic research on how language-model agents learn, interact, and generalize.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.34214 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T20:29:01.481Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).