Metraipolis › archive › dc_ca43cb2b7b7f
Consolidating RLVR Capabilities Across Domains: A Deep Dive into Fusion Paradigms — arXiv:2608.27409
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.27409
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T15:20:45.831Z
- SHA-256 of the archived text
ef99b63a73ac1f89ea06c003c9367b490e249b6e0c1ad5d984836379874723f7- Stable URL of this page
https://metraipolis.com/documento/dc_ca43cb2b7b7f
How to cite
Consolidating RLVR Capabilities Across Domains: A Deep Dive into Fusion Paradigms — arXiv:2608.27409 — arXiv (en), https://arxiv.org/abs/2608.27409 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ca43cb2b7b7f (Metraipolis archive dc_ca43cb2b7b7f, sha256 ef99b63a73ac1f89ea06c003c9367b490e249b6e0c1ad5d984836379874723f7)
BibTeX
@misc{dcca43cb2b7b7f,
title = {Consolidating RLVR Capabilities Across Domains: A Deep Dive into Fusion Paradigms — arXiv:2608.27409},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.27409},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ca43cb2b7b7f; sha256 ef99b63a73ac1f89ea06c003c9367b490e249b6e0c1ad5d984836379874723f7},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.27409. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Reinforcement learning with verifiable rewards (RLVR) improves specific capabilities of large language models, but covering multiple capabilities often involves training separate domain experts and subsequently consolidating them. We organize three fusion paradigms by the artifacts they reuse: Merge combines expert task vectors, Mix RL pools their datasets, and multi-teacher on-policy distillation (MOPD) uses both. Because they have largely been studied in isolation, how they compare and how to choose among them remain unclear. We compare all three using shared experts and data across model scales and a multi-domain benchmark suite. Although their average performance differs by at most 1.4 points, the gap reaches 8.6 points on a single benchmark, with domain-level variation tracking cross-domain relations visible in task-vector geometry. Training dynamics expose distinct constraints: Mix RL depends on domain mixture proportions, MOPD remains bounded by its teachers, and Merge compresses all expert updates into one. All three improve single-sample accuracy without measurable gains in solution coverage or losses in held-out capabilities. These results yield a practical guideline: use Merge when experts already exist and cheap fusion is paramount; Mix RL when training a unified model without experts, with domain proportions adjusted for cross-domain transfer; and MOPD when preserving domain-specific gains matters more than surpassing teachers or minimizing end-to-end cost. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-28 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2608.27409 — 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 2026-10-17T15:15:18.100Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).