Metraipolis › archive › dc_e7b48477e73c

StudyBench: Can Self-Evolution Squeeze Textbooks for Olympiad Capability? — arXiv:2609.00787

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.00787
License (as archived)
CC0-1.0
Archived
2026-10-17T06:32:07.203Z
SHA-256 of the archived text
dca0d0b5e83a127e05dc9625fc4bd1bc2d33da73c9a2e16054ced6741d5b7f3f
Stable URL of this page
https://metraipolis.com/documento/dc_e7b48477e73c

How to cite

StudyBench: Can Self-Evolution Squeeze Textbooks for Olympiad Capability? — arXiv:2609.00787 — arXiv (en), https://arxiv.org/abs/2609.00787 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e7b48477e73c (Metraipolis archive dc_e7b48477e73c, sha256 dca0d0b5e83a127e05dc9625fc4bd1bc2d33da73c9a2e16054ced6741d5b7f3f)
BibTeX
@misc{dce7b48477e73c,
  title = {StudyBench: Can Self-Evolution Squeeze Textbooks for Olympiad Capability? — arXiv:2609.00787},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.00787},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e7b48477e73c; sha256 dca0d0b5e83a127e05dc9625fc4bd1bc2d33da73c9a2e16054ced6741d5b7f3f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.00787. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Humans need to study only a handful of well-written textbooks to master a discipline and attempt its hardest problems. We argue that an ideal self-evolution method should share the same property, that is autonomously learning from raw training material for transferable problem-solving capability. However, we still lack a direct measurement for it. We introduce StudyBench, a controlled physics benchmark that directly measures how efficiently a self-evolution method converts training material into capability. We organise the test set into an Application Set, consisting of difficult textbook problems and evaluating absorption ability, and a Transfer Set, consisting of olympiad-level problems and evaluating transfer ability. Benchmarking representative self-evolution methods across three base models, we find that improvements on the Application Set rarely translate to the harder Transfer Set. A guidance ablation exposes a Guidance Gap: even the strongest method closes only a small fraction of what the same material unlocks when supplied as in-context guidance. Besides, every method hits a Compute Plateau, saturating well before exhausting its compute budget. The remaining gap is therefore a method problem rather than a data or compute problem. By offering a clean and controlled benchmark, StudyBench turns self-evolution progress from an open-ended pursuit into a measurable target for future research. Our code is released at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.00787 — 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 2026-10-17T06:15:12.643Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).