paper-with-me

홈 › Papers

Sequence-Level Certainty Reduces Hallucination In Knowledge-Grounded Dialogue Generation

2023-10-28 · Yixin Wan, Fanyou Wu, Weijie Xu, Srinivasan H. Sengamedu

In this work, we propose sequence-level certainty as a common theme over hallucination in Knowledge Grounded Dialogue Generation (KGDG). We explore the correlation between the level of hallucination in model responses and two types of sequence-level certainty: probabilistic certainty and semantic certainty. Empirical results reveal that higher levels of both types of certainty in model responses are correlated with lower levels of hallucination. We further propose Certainty-based Response Ranking (CRR), a decoding-time hallucination mitigation method that samples several response candidates, ranks them based on sequence-level certainty, and outputs the response with the highest certainty level. Aligning with our definitions of sequence-level certainty, we design 2 types of CRR approaches: Probabilistic CRR (P-CRR) and Semantic CRR (S-CRR). P-CRR ranks individually sampled model responses using the arithmetic mean log-probability of the entire sequence. S-CRR approaches certainty estimation from meaning-space, and ranks model response candidates based on their semantic certainty level as measured by an entailment-based Agreement Score (AS). Through extensive experiments across 3 KGDG datasets, 3 decoding methods, and 4 KGDG models, we validate the effectiveness of CRR for reducing hallucination in KGDG task.

📄 PDF Abstract BibTeX arXiv:2310.18794

Code (0)

등록된 구현이 없습니다.

Tasks

Dialogue GenerationHallucinationText Generation

Similar Papers 제목 키워드 기반

Chain-based Adaptive Reconfiguration Over Lattices for Hallucination Reduction

2026-05-26 · Joan Vendrell Gallart, Solmaz Kia, Russell Bent, Michael Grosskopf arxiv

We introduce CAROL (Chain-based Adaptive Reconfiguration Over Lattices), a probabilistic framework for test-time hallucination reduction in large language models. Rather than relying on token-level uncertainty, CAROL def…

Computational EfficiencyQuestion Answering

CLUE: Concept-Level Uncertainty Estimation for Large Language Models

2024-09-04 · Yu-Hsiang Wang, Andrew Bai, Che-Ping Tsai, Cho-Jui Hsieh

Large Language Models (LLMs) have demonstrated remarkable proficiency in various natural language generation (NLG) tasks. Previous studies suggest that LLMs' generation process involves uncertainty. However, existing app…

HallucinationSentenceStory GenerationText Generation

INSIDE: LLMs' Internal States Retain the Power of Hallucination Detection

2024-02-06 · Chao Chen, Kai Liu, Ze Chen, Yi Gu 외

Knowledge hallucination have raised widespread concerns for the security and reliability of deployed LLMs. Previous efforts in detecting hallucinations have been employed at logit-level uncertainty estimation or language…

DiversityHallucinationQuestion Answering

Enhancing Uncertainty Modeling with Semantic Graph for Hallucination Detection

2025-01-02 · Kedi Chen, Qin Chen, Jie zhou, Xinqi Tao 외

Large Language Models (LLMs) are prone to hallucination with non-factual or unfaithful statements, which undermines the applications in real-world scenarios. Recent researches focus on uncertainty-based hallucination det…

HallucinationSentence

Smoothing Out Hallucinations: Mitigating LLM Hallucination with Smoothed Knowledge Distillation

2025-02-16 · Hieu Nguyen, Zihao He, Shoumik Atul Gandre, Ujjwal Pasupulety 외

Large language models (LLMs) often suffer from hallucination, generating factually incorrect or ungrounded content, which limits their reliability in high-stakes applications. A key factor contributing to hallucination i…

HallucinationKnowledge Distillation