Metraipolis › archive › dc_32b4fba57cd3
SFAD: Speculative Factuality-Aware Decoding — arXiv:2609.00796
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.00796
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T06:41:22.676Z
- SHA-256 of the archived text
f5e4760e9813dfd70452f082628b7a1151ccab275bb732214303683bd6e9fcef- Stable URL of this page
https://metraipolis.com/documento/dc_32b4fba57cd3
How to cite
SFAD: Speculative Factuality-Aware Decoding — arXiv:2609.00796 — arXiv (en), https://arxiv.org/abs/2609.00796 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_32b4fba57cd3 (Metraipolis archive dc_32b4fba57cd3, sha256 f5e4760e9813dfd70452f082628b7a1151ccab275bb732214303683bd6e9fcef)
BibTeX
@misc{dc32b4fba57cd3,
title = {SFAD: Speculative Factuality-Aware Decoding — arXiv:2609.00796},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.00796},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_32b4fba57cd3; sha256 f5e4760e9813dfd70452f082628b7a1151ccab275bb732214303683bd6e9fcef},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.00796. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
As one of the most critical challenges in large language models, contextual faithfulness directly determines their reliability in knowledge-intensive applications. This task is particularly challenging as it requires balancing factual consistency with generation efficiency. Contrastive decoding methods require dual forward passes (with and without context) to compare model outputs, doubling inference computational overhead, while post-training alignment demands extensive reinforcement learning with substantial computational overhead. To address this challenge, we present SFAD, a speculative decoding framework that enhances contextual faithfulness without inference degradation. We first construct ConFide, a preference dataset with fine-grained atomic perturbations, to train a context-faithful draft model via Direct Preference Optimization. During inference, Epistemic Friction detects potential hallucinations by quantifying distributional tension weighted by specialist certainty. When friction exceeds the threshold, Asymmetric Logit Steering refines the target distribution through residual-based logit injection; otherwise, standard speculation proceeds. Extensive experiments demonstrate that SFAD substantially improves faithfulness while achieving $2.48\times$ speedup, offering a practical solution for efficient LLMs. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.00796 — 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 2026-10-17T06:33:38.828Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).