paper-with-me

홈 › Papers

How Decoding Strategies Affect the Verifiability of Generated Text

2019-11-09 · Findings of the Association for Computational Linguistics 2020 · Luca Massarelli, Fabio Petroni, Aleksandra Piktus, Myle Ott, Tim Rocktäschel, Vassilis Plachouras, Fabrizio Silvestri, Sebastian Riedel

Recent progress in pre-trained language models led to systems that are able to generate text of an increasingly high quality. While several works have investigated the fluency and grammatical correctness of such models, it is still unclear to which extent the generated text is consistent with factual world knowledge. Here, we go beyond fluency and also investigate the verifiability of text generated by state-of-the-art pre-trained language models. A generated sentence is verifiable if it can be corroborated or disproved by Wikipedia, and we find that the verifiability of generated text strongly depends on the decoding strategy. In particular, we discover a tradeoff between factuality (i.e., the ability of generating Wikipedia corroborated text) and repetitiveness. While decoding strategies such as top-k and nucleus sampling lead to less repetitive generations, they also produce less verifiable text. Based on these finding, we introduce a simple and effective decoding strategy which, in comparison to previously used decoding strategies, produces less repetitive and more verifiable text.

📄 PDF Abstract BibTeX arXiv:1911.03587

Code (1)

TaiseiAso/BiGruAttEncDec pytorch

Tasks

Language ModellingNatural Language UnderstandingSentenceWorld Knowledge

Similar Papers 제목 키워드 기반

Energy-Conscious LLM Decoding: Impact of Text Generation Strategies on GPU Energy Consumption

2025-02-17 · Alireza Nik, Michael A. Riegler, Pål Halvorsen

Decoding strategies significantly influence the quality and diversity of the generated texts in large language models (LLMs), yet their impact on computational resource consumption, particularly GPU energy usage, is insu…

BenchmarkingCode SummarizationDiversityGPU+2

How Sampling Affects the Detectability of Machine-written texts: A Comprehensive Study

2025-10-15 · Matthieu Dubois, François Yvon, Pablo Piantanida arxiv

As texts generated by Large Language Models (LLMs) are ever more common and often indistinguishable from human-written content, research on automatic text detection has attracted growing attention. Many recent detectors …

Text Detection

Assessing the Verifiability of Attributions in News Text

2017-11-01 · IJCNLP 2017 11 · Edward Newell, Ariane Schang, Drew Margolin, Derek Ruths

When reporting the news, journalists rely on the statements of stakeholders, experts, and officials. The attribution of such a statement is verifiable if its fidelity to the source can be confirmed or denied. In this pap…

Question Answering

Diversifying to Verify: When Task-Equivalent Programs Differ in Verifiability

2026-07-10 · Shirley Yu, Ruben Martins arxiv

Program verification is crucial for software correctness, but producing fully verified programs remains difficult in practice. This paper studies whether implementation structure affects automated verifiability when mult…

ClearSight: Visual Signal Enhancement for Object Hallucination Mitigation in Multimodal Large language Models

2025-03-17 · CVPR 2025 1 · Hao Yin, Guangzong Si, Zilei Wang

Contrastive decoding strategies are widely used to mitigate object hallucinations in multimodal large language models (MLLMs). By reducing over-reliance on language priors, these strategies ensure that generated content …

Computational EfficiencyHallucinationObject Hallucination