Metraipolis › archive › dc_91b18fdf352c

CritICL: Inference-Time Weak-to-Strong Generalization from Small Language Model Failure Modes — arXiv:2608.27455

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.27455
License (as archived)
CC0-1.0
Archived
2026-10-17T15:20:45.907Z
SHA-256 of the archived text
525272b0ea25c3d5193e70f1f044ac4e1615e6d3f957e1c54b100f59f6f05a44
Stable URL of this page
https://metraipolis.com/documento/dc_91b18fdf352c

How to cite

CritICL: Inference-Time Weak-to-Strong Generalization from Small Language Model Failure Modes — arXiv:2608.27455 — arXiv (en), https://arxiv.org/abs/2608.27455 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_91b18fdf352c (Metraipolis archive dc_91b18fdf352c, sha256 525272b0ea25c3d5193e70f1f044ac4e1615e6d3f957e1c54b100f59f6f05a44)
BibTeX
@misc{dc91b18fdf352c,
  title = {CritICL: Inference-Time Weak-to-Strong Generalization from Small Language Model Failure Modes — arXiv:2608.27455},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.27455},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_91b18fdf352c; sha256 525272b0ea25c3d5193e70f1f044ac4e1615e6d3f957e1c54b100f59f6f05a44},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.27455. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Recent advances in inference-time scaling have significantly improved the reasoning performance of large language models (LLMs). However, these methods typically rely on repeated generation or external verification. To address this limitation, we introduce CritICL, a novel inference-time framework that improves reasoning while maintaining high efficiency. Our key insight is that LLM failure modes exhibit structured patterns across model scales within the same family. Instead of treating failures as undesirable outputs, CritICL leverages them as a source of guidance. Specifically, we utilize failure modes derived from weaker models and incorporate them into inference through critique-based in-context examples. We propose two variants: CritICL-dynamic, which adaptively predicts input-specific failure modes and retrieves critiques, and CritICL-static, which uses a global failure mode profile to provide stable guidance. Experimental results show that CritICL consistently outperforms standard in-context learning and achieves performance competitive with or superior to test-time scaling methods, while requiring significantly fewer generations and lower token cost. Code available at: this https URL

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/2608.27455 — 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 2026-10-17T15:15:18.104Z 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).