# ACL-Verbatim: hallucination-free question answering for research

arXiv: https://arxiv.org/abs/2605.21102 · HTML: https://mlsanity.com/paper/2605.21102/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 55 references that are in the map.
- First posted on arXiv 20 May 2026.

## Builds on (most cited first, 20 of 20)

- [SciBERT: A Pretrained Language Model for Scientific Text](https://mlsanity.com/paper/1903.10676/) (2019)
- [Galactica: A Large Language Model for Science](https://mlsanity.com/paper/2211.09085/) (2022)
- [ELI5: Long Form Question Answering](https://mlsanity.com/paper/1907.09190/) (2019)
- [A Dataset of Information-Seeking Questions and Answers Anchored in Research Papers](https://mlsanity.com/paper/2105.03011/) (2021)
- [RAGTruth: A Hallucination Corpus for Developing Trustworthy Retrieval-Augmented Language Models](https://mlsanity.com/paper/2401.00396/) (2023)
- [ARES: An Automated Evaluation Framework for Retrieval-Augmented Generation Systems](https://mlsanity.com/paper/2311.09476/) (2023)
- [ASQA: Factoid Questions Meet Long-Form Answers](https://mlsanity.com/paper/2204.06092/) (2022)
- [Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Tools](https://mlsanity.com/paper/2405.20362/) (2024)
- [RAGBench: Explainable Benchmark for Retrieval-Augmented Generation Systems](https://mlsanity.com/paper/2407.11005/) (2024)
- [Are self-explanations from Large Language Models faithful?](https://mlsanity.com/paper/2401.07927/) (2024)
- [Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanations](https://mlsanity.com/paper/2307.08678/) (2023)
- [Chainpoll: A high efficacy method for LLM hallucination detection](https://mlsanity.com/paper/2310.18344/) (2023)
- [Clinfo.ai: An Open-Source Retrieval-Augmented Large Language Model System for Answering Medical Questions using Scientific Literature](https://mlsanity.com/paper/2310.16146/) (2023)
- [CLAPNQ: Cohesive Long-form Answers from Passages in Natural Questions for RAG systems](https://mlsanity.com/paper/2404.02103/) (2024)
- [The ACL OCL Corpus: Advancing Open Science in Computational Linguistics](https://mlsanity.com/paper/2305.14996/) (2023)
- [DataFinder: Scientific Dataset Recommendation from Natural Language Descriptions](https://mlsanity.com/paper/2305.16636/) (2023)
- [FACTOID: FACtual enTailment fOr hallucInation Detection](https://mlsanity.com/paper/2403.19113/) (2024)
- [NLPExplorer: Exploring the Universe of NLP Papers](https://mlsanity.com/paper/1910.07351/) (2019)
- [ScIRGen: Synthesize Realistic and Large-Scale RAG Dataset for Scientific Research](https://mlsanity.com/paper/2506.11117/) (2025)
- [Squeez: Task-Conditioned Tool-Output Pruning for Coding Agents](https://mlsanity.com/paper/2604.04979/) (2026)

## Built on by (most cited first, 1 of 1)

- [Beyond Document Grounding: Span-Level Hallucination Detection over Code, Tool Output, and Documents](https://mlsanity.com/paper/2607.00895/) (2026)

## Abstract (arXiv)

Academic researchers need efficient and reliable methods for collecting high-quality information from trusted sources, but modern tools for AI-assisted research still suffer from the tendency of Large Language Models (LLMs) to produce factually inaccurate or nonsensical output, commonly referred to as hallucinations. We apply the extractive question answering system VerbatimRAG to research papers in the ACL Anthology, directly mapping user queries to verbatim text spans in retrieved documents. We contribute a novel ground truth dataset for the task of mapping user queries to relevant text spans in research papers, and use it to train and evaluate a variety of extractive models. Human annotation is performed by NLP researchers and is based on synthetic user queries generated using a custom pipeline based on the ScIRGen methodology, paired with chunks of research papers retrieved by VerbatimRAG. On this benchmark, a 150M-parameter ModernBERT token classifier trained on silver supervision from our pipeline achieves the best word-level F1 (53.6), ahead of the strongest evaluated LLM extractor (48.7).