Metraipolis › archive › dc_55de46aa22f7
Primitives and Paths: Predicting What Post-Training Can Do from What the Base Model Already Has — arXiv:2611.11260
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.11260
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-10T18:43:15.225Z
- SHA-256 of the archived text
1ee9b38fb4dc7a7619407e1052dd9fd45aa4af28a066cb6f9044dd0161fedfb1- Stable URL of this page
https://metraipolis.com/documento/dc_55de46aa22f7
How to cite
Primitives and Paths: Predicting What Post-Training Can Do from What the Base Model Already Has — arXiv:2611.11260 — arXiv (en), https://arxiv.org/abs/2611.11260 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_55de46aa22f7 (Metraipolis archive dc_55de46aa22f7, sha256 1ee9b38fb4dc7a7619407e1052dd9fd45aa4af28a066cb6f9044dd0161fedfb1)
BibTeX
@misc{dc55de46aa22f7,
title = {Primitives and Paths: Predicting What Post-Training Can Do from What the Base Model Already Has — arXiv:2611.11260},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.11260},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_55de46aa22f7; sha256 1ee9b38fb4dc7a7619407e1052dd9fd45aa4af28a066cb6f9044dd0161fedfb1},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.11260. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Does reinforcement learning teach language models new capabilities, or merely amplify ones acquired during pre-training? The literature offers evidence for both: pass@$k$ studies suggest RL reweights latent behaviours while some controlled composition studies demonstrate gains beyond simple amplification. We argue that this question has no answer independent of the capability state of the base model, and propose a two-dimensional state space for post-trainability. The first axis, \emph{effective full-solution support}, measures whether successful trajectories can be sampled under a feasible exploration budget. The second axis, \emph{atomic primitive coverage}, measures whether the constituent skills of a solution can be executed when the burden of composition is removed. Together, these dimensions motivate three regimes -- amplification, composition-limited, and acquisition-limited -- each predicting a distinct empirical signature and calling for a distinct intervention. We propose a prospective measurement protocol and investigate the framework using synthetic composition tasks across three small language models. We find that primitive mastery alone does not guarantee that RL can learn a composition. Checkpoints that execute every primitive reliably can fail to improve when complete solutions are never sampled during training, whereas modest composition supervision that establishes nonzero support enables substantial RL gains. The framework reframes mid-training as preparation for post-training: its role is to move models into states from which RL can operate effectively.
Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.11260 — 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 2026-11-10T18:37:37.632Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).