DeepAI AI Chat
Log In Sign Up

A Token-level Reference-free Hallucination Detection Benchmark for Free-form Text Generation

04/18/2021
by   Tianyu Liu, et al.
0

Large pretrained generative models like GPT-3 often suffer from hallucinating non-existent or incorrect content, which undermines their potential merits in real applications. Existing work usually attempts to detect these hallucinations based on a corresponding oracle reference at a sentence or document level. However ground-truth references may not be readily available for many free-form text generation applications, and sentence- or document-level detection may fail to provide the fine-grained signals that would prevent fallacious content in real time. As a first step to addressing these issues, we propose a novel token-level, reference-free hallucination detection task and an associated annotated dataset named HaDes (HAllucination DEtection dataSet). To create this dataset, we first perturb a large number of text segments extracted from English language Wikipedia, and then verify these with crowd-sourced annotations. To mitigate label imbalance during annotation, we utilize an iterative model-in-loop strategy. We conduct comprehensive data analyses and create multiple baseline models.

READ FULL TEXT

page 5

page 6

05/13/2019

Towards Content Transfer through Grounded Text Generation

Recent work in neural generation has attracted significant interest in c...
10/07/2020

Detecting Fine-Grained Cross-Lingual Semantic Divergences without Supervision by Learning to Rank

Detecting fine-grained differences in content conveyed in different lang...
03/23/2017

TokTrack: A Complete Token Provenance and Change Tracking Dataset for the English Wikipedia

We present a dataset that contains every instance of all tokens ( words...
08/06/2021

Sentence Semantic Regression for Text Generation

Recall the classical text generation works, the generation framework can...
10/31/2022

Semantic Novelty Detection and Characterization in Factual Text Involving Named Entities

Much of the existing work on text novelty detection has been studied at ...
01/11/2023

Diving Deep into Modes of Fact Hallucinations in Dialogue Systems

Knowledge Graph(KG) grounded conversations often use large pre-trained m...
01/20/2021

Generating (Formulaic) Text by Splicing Together Nearest Neighbors

We propose to tackle conditional text generation tasks, especially those...