Metraipolis › archive › dc_8a7cef228ee0

TsuGO: Probing Search Efficiency in LLM Reasoning via Go Life-and-Death Problems — arXiv:2608.13221

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.13221
License (as archived)
CC0-1.0
Archived
2026-10-19T04:17:35.647Z
SHA-256 of the archived text
9bdae32e46b25450bc4399a7892034a96620bdc2dcbc30f823baade2f5716107
Stable URL of this page
https://metraipolis.com/documento/dc_8a7cef228ee0

How to cite

TsuGO: Probing Search Efficiency in LLM Reasoning via Go Life-and-Death Problems — arXiv:2608.13221 — arXiv (en), https://arxiv.org/abs/2608.13221 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_8a7cef228ee0 (Metraipolis archive dc_8a7cef228ee0, sha256 9bdae32e46b25450bc4399a7892034a96620bdc2dcbc30f823baade2f5716107)
BibTeX
@misc{dc8a7cef228ee0,
  title = {TsuGO: Probing Search Efficiency in LLM Reasoning via Go Life-and-Death Problems — arXiv:2608.13221},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.13221},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_8a7cef228ee0; sha256 9bdae32e46b25450bc4399a7892034a96620bdc2dcbc30f823baade2f5716107},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.13221. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

The evaluation of LLM reasoning is moving from final-answer accuracy to process-level assessment, yet existing methods still fail to capture how models plan reasoning paths and allocate reasoning resources--that is, how they organize search. Prior process-level methods focus on the coherence and redundancy of chain-of-thought (CoT), and most benchmark tasks have a single objective solvable by static capabilities such as derivation and tool use, leaving search organization unmeasured. We introduce TsuGO, a process-level reasoning benchmark for evaluating Search Efficiency in LLM reasoning through Go life-and-death problems. These problems provide closed and verifiable solution spaces with an inherent adversarial structure, making candidate generation, response checking, branch comparison, and backtracking necessary parts of reasoning rather than incidental trace patterns. By constraining the solution space, TsuGO disentangles domain knowledge from search organization, parses CoT into a structured search tree, and reports Search Efficiency together with Token Efficiency and other diagnostic metrics and visualizations. Experiments show that current LLMs remain far from stable tsumego solving: stronger models succeed by finding the correct candidate earlier and sustaining effort on productive branches, but most models still behave much closer to unguided search algorithms than to neural-guided KataGo. Longer CoT or higher Token Efficiency does not necessarily imply better search. Our results identify search organization and reasoning-resource allocation as missing dimensions in LLM reasoning evaluation.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-14 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.13221 — 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 2026-10-19T04:13:06.591Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).