Metraipolis › archive › dc_56ca02ead384

CausalOPD: First-Wrong-Step Supervision for Distilling Causal Chain Reasoning — arXiv:2608.03673

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.03673
License (as archived)
CC0-1.0
Archived
2026-10-31T12:42:29.317Z
SHA-256 of the archived text
6180ad0b060724f8c88287bcfafb8b8a4769689636490ccf6a1c89edbaabc2a8
Stable URL of this page
https://metraipolis.com/documento/dc_56ca02ead384

How to cite

CausalOPD: First-Wrong-Step Supervision for Distilling Causal Chain Reasoning — arXiv:2608.03673 — arXiv (en), https://arxiv.org/abs/2608.03673 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_56ca02ead384 (Metraipolis archive dc_56ca02ead384, sha256 6180ad0b060724f8c88287bcfafb8b8a4769689636490ccf6a1c89edbaabc2a8)
BibTeX
@misc{dc56ca02ead384,
  title = {CausalOPD: First-Wrong-Step Supervision for Distilling Causal Chain Reasoning — arXiv:2608.03673},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.03673},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_56ca02ead384; sha256 6180ad0b060724f8c88287bcfafb8b8a4769689636490ccf6a1c89edbaabc2a8},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.03673. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Many critical reasoning tasks, including clinical diagnosis, legal judgment, and industrial fault diagnosis, require step-dependent causal chains in which early errors propagate and correct conclusions can mask invalid reasoning. Although large language models perform well on such tasks, privacy, latency, and controllability motivate distillation into locally deployable models. Standard trajectory imitation does not correct process errors on the student's own rollout distribution. We propose CausalOPD, a curriculum online process distillation framework. A knowledge-augmented teacher first provides trajectories grounded in domain-specific causal rules, entity relations, and structural constraints. The student then generates on-policy trajectories, and the teacher identifies the first wrong step, defined as the earliest transition that verifiably violates available constraints. Starting from the verified prefix, short-horizon reinforcement learning repairs this localized failure. A causal-stage curriculum advances from evidence-level to mechanism-level and conclusion-level errors, following their propagation order. Across three domains, CausalOPD improves average path correctness by 23.4 percentage points over sequence-level online process distillation and reduces the right-label-wrong-reasoning rate from 15.7% to 4.4%. The domain-specific 8B students also surpass both evaluated proprietary references in path correctness across all domains.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-05 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2608.03673 — 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 2026-10-31T12:31:30.363Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-08-05?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).