Metraipolis › archive › dc_4b99662453bd

RLVR Sharpens the Pencil, Not the Mind — arXiv:2611.02574

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.02574
License (as archived)
CC0-1.0
Archived
2026-11-03T18:37:24.418Z
SHA-256 of the archived text
fd1b5e157b3f2cf48e7cb90d429824eb7ca0b4ac1b6d2db1743ba2d4bbd3ecd1
Stable URL of this page
https://metraipolis.com/documento/dc_4b99662453bd

How to cite

RLVR Sharpens the Pencil, Not the Mind — arXiv:2611.02574 — arXiv (en), https://arxiv.org/abs/2611.02574 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4b99662453bd (Metraipolis archive dc_4b99662453bd, sha256 fd1b5e157b3f2cf48e7cb90d429824eb7ca0b4ac1b6d2db1743ba2d4bbd3ecd1)
BibTeX
@misc{dc4b99662453bd,
  title = {RLVR Sharpens the Pencil, Not the Mind — arXiv:2611.02574},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.02574},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4b99662453bd; sha256 fd1b5e157b3f2cf48e7cb90d429824eb7ca0b4ac1b6d2db1743ba2d4bbd3ecd1},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.02574. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Models like DeepSeek-R1 made Reinforcement Learning with Verifiable Rewards (RLVR) famous: reward a model for correct answers and it gets dramatically better at reasoning without being taught how. The leading explanation is sharpening---that the base model already knows how to solve these problems and RLVR just makes it more likely to reach for its best solution paths. This has only ever been checked by comparing before and after training, never by watching what changes inside across it. We tracked OLMo-2-1B across all 13 training checkpoints on GSM8K. First-try accuracy climbs from 58.0% to 61.3%, but ask the model 256 times and solution coverage actually rises from 94.5% to 96.9%, proving the knowledge was already there. We then planted reasoning mistakes into problems and trained simple probes to detect them. Those probes held between AUROC 0.798 and 0.840 across all 13 checkpoints and 3 error types, a spread smaller than the confidence interval; the model never builds a sharper internal picture of what a reasoning mistake looks like. The one place we find real internal change is at the final answer token, where a probe distinguishing early from late checkpoints hits an AUROC of 0.93, while the same probe on the actual reasoning steps lands near chance. RLVR makes the model lead with its best thinking. It does not change what that thinking is.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2611.02574 — 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 2026-11-03T18:12:54.578Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).