Metraipolis › archive › dc_98c100b22440
Is Discrete Difficulty Sufficient? Leveraging Continuous Difficulty for Efficient Self-Consistency in LLMs — arXiv:2608.24590
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.24590
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T06:18:08.225Z
- SHA-256 of the archived text
bbfdc743877f6be24260b51b0d9cfbe5400d4fde748eea31dcc672667bcf35d8- Stable URL of this page
https://metraipolis.com/documento/dc_98c100b22440
How to cite
Is Discrete Difficulty Sufficient? Leveraging Continuous Difficulty for Efficient Self-Consistency in LLMs — arXiv:2608.24590 — arXiv (en), https://arxiv.org/abs/2608.24590 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_98c100b22440 (Metraipolis archive dc_98c100b22440, sha256 bbfdc743877f6be24260b51b0d9cfbe5400d4fde748eea31dcc672667bcf35d8)
BibTeX
@misc{dc98c100b22440,
title = {Is Discrete Difficulty Sufficient? Leveraging Continuous Difficulty for Efficient Self-Consistency in LLMs — arXiv:2608.24590},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.24590},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_98c100b22440; sha256 bbfdc743877f6be24260b51b0d9cfbe5400d4fde748eea31dcc672667bcf35d8},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.24590. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Self-Consistency (SC) is a decoding strategy that samples diverse reasoning paths and selects the most consistent answer, demonstrating strong performance on complex reasoning problems. However, the excessive token consumption incurred by generating multiple reasoning paths has been identified as a major limitation of SC. To improve computational efficiency, several studies have proposed strategies that adjust the number of reasoning paths or allocate resources differentially according to problem difficulty. Nevertheless, most existing methods categorize difficulty into a few fixed levels, failing to fully capture the continuously varying nature of reasoning complexity. In this work, we propose Flexible Self-Consistency (FSC), which estimates problem difficulty as a continuous signal and dynamically adjusts the number of generated reasoning paths accordingly. FSC predicts the output entropy of an input question using a pre-trained probe and leverages it as an indicator of model uncertainty to flexibly control the sampling budget. Experimental results show that, across various models and benchmarks, FSC maintains accuracy comparable to SC while achieving token savings of up to 76%. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-26 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2608.24590 — 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 2026-10-18T06:17:29.371Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-26?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).