# Think While You Write: Hypothesis Verification Promotes Faithful Knowledge-to-Text Generation

arXiv: https://arxiv.org/abs/2311.09467 · HTML: https://mlsanity.com/paper/2311.09467/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 41 of its 53 references that are in the map.
- First posted on arXiv 16 Nov 2023.

## Builds on (most cited first, 30 of 41)

- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://mlsanity.com/paper/1910.13461/) (2019)
- [BERTScore: Evaluating Text Generation with BERT](https://mlsanity.com/paper/1904.09675/) (2019)
- [Adversarial NLI: A New Benchmark for Natural Language Understanding](https://mlsanity.com/paper/1910.14599/) (2019)
- [Evaluating the Factual Consistency of Abstractive Text Summarization](https://mlsanity.com/paper/1910.12840/) (2019)
- [SummaC: Re-Visiting NLI-based Models for Inconsistency Detection in Summarization](https://mlsanity.com/paper/2111.09525/) (2021)
- [How Language Model Hallucinations Can Snowball](https://mlsanity.com/paper/2305.13534/) (2023)
- [On Hallucination and Predictive Uncertainty in Conditional Language Generation](https://mlsanity.com/paper/2103.15025/) (2021)
- [BRIO: Bringing Order to Abstractive Summarization](https://mlsanity.com/paper/2203.16804/) (2022)
- [NeuroLogic A*esque Decoding: Constrained Text Generation with Lookahead Heuristics](https://mlsanity.com/paper/2112.08726/) (2021)
- [Generalization in Generation: A closer look at Exposure Bias](https://mlsanity.com/paper/1910.00292/) (2019)
- [Text Generation from Knowledge Graphs with Graph Transformers](https://mlsanity.com/paper/1904.02342/) (2019)
- [Investigating Pretrained Language Models for Graph-to-Text Generation](https://mlsanity.com/paper/2007.08426/) (2020)
- [FaithDial: A Faithful Benchmark for Information-Seeking Dialogue](https://mlsanity.com/paper/2204.10757/) (2022)
- [Data-to-Text Generation with Content Selection and Planning](https://mlsanity.com/paper/1809.00582/) (2018)
- [TableFormer: Robust Transformer Modeling for Table-Text Encoding](https://mlsanity.com/paper/2203.00274/) (2022)
- [FactKB: Generalizable Factuality Evaluation using Language Models Enhanced with Factual Knowledge](https://mlsanity.com/paper/2305.08281/) (2023)
- [Detecting and Mitigating Hallucinations in Multilingual Summarisation](https://mlsanity.com/paper/2305.13632/) (2023)
- [RHO (ρ): Reducing Hallucination in Open-domain Dialogues with Knowledge Grounding](https://mlsanity.com/paper/2212.01588/) (2022)
- [Understanding and Detecting Hallucinations in Neural Machine Translation via Model Introspection](https://mlsanity.com/paper/2301.07779/) (2023)
- [UniRE: A Unified Label Space for Entity Relation Extraction](https://mlsanity.com/paper/2107.04292/) (2021)
- [Data-to-text Generation with Macro Planning](https://mlsanity.com/paper/2102.02723/) (2021)
- [Falsesum: Generating Document-level NLI Examples for Recognizing Factual Inconsistency in Summarization](https://mlsanity.com/paper/2205.06009/) (2022)
- [Robust (Controlled) Table-to-Text Generation with Structure-Aware Equivariance Learning](https://mlsanity.com/paper/2205.03972/) (2022)
- [Commonsense-Focused Dialogues for Response Generation: An Empirical Study](https://mlsanity.com/paper/2109.06427/) (2021)
- [Densely Connected Graph Convolutional Networks for Graph-to-Sequence Learning](https://mlsanity.com/paper/1908.05957/) (2019)
- [Few-shot Knowledge Graph-to-Text Generation with Pretrained Language Models](https://mlsanity.com/paper/2106.01623/) (2021)
- [Elastic Weight Removal for Faithful and Abstractive Dialogue Generation](https://mlsanity.com/paper/2303.17574/) (2023)
- [Deep Graph Convolutional Encoders for Structured Data to Text Generation](https://mlsanity.com/paper/1810.09995/) (2018)

## Built on by (most cited first, 5 of 5)

- [A Comprehensive Survey of Hallucination Mitigation Techniques in Large Language Models](https://mlsanity.com/paper/2401.01313/) (2024)
- [FACTOID: FACtual enTailment fOr hallucInation Detection](https://mlsanity.com/paper/2403.19113/) (2024)
- [HalluClean: A Unified Framework to Combat Hallucinations in LLMs](https://mlsanity.com/paper/2511.08916/) (2025)
- [Multilingual Fine-Grained News Headline Hallucination Detection](https://mlsanity.com/paper/2407.15975/) (2024)
- [Attribution Techniques for Mitigating Hallucinated Information in RAG Systems: A Survey](https://mlsanity.com/paper/2601.19927/) (2026)

## Abstract (arXiv)

Knowledge-to-text generators often struggle to faithfully generate descriptions for the input facts: they may produce hallucinations that contradict the input, or describe facts not present in the input. To reduce hallucinations, we propose a decoding-only method, TWEAK (Think While Effectively Articulating Knowledge), which can be integrated with any generator without retraining. TWEAK treats the generated sequences at each decoding step and its future sequences as hypotheses, and ranks each generation candidate based on the extent to which their hypotheses are supported by the input facts using a Hypothesis Verification Model (HVM). We first demonstrate the effectiveness of TWEAK by using a Natural Language Inference (NLI) model as the HVM and report improved faithfulness with a minimal impact on the quality. We then replace the NLI model with a task-specific HVM trained with a first-of-a-kind dataset, FATE (Fact-Aligned Textual Entailment), which pairs input facts with their original and perturbed descriptions. We test TWEAK with two generators, and the best TWEAK variants improve on average for the two models by 2.24/7.17 points in faithfulness (FactKB) in in/out-of-distribution evaluations, respectively, and with only a 0.14/0.32-point decline in quality (BERTScore).