HackerPapers

arXiv:2305.13534cs.CL

How Language Model Hallucinations Can Snowball

Muru Zhang, Ofir Press, William Merrill, Alisa Liu, Noah A. Smith

Abstract

A major risk of using language models in practical applications is their tendency to hallucinate incorrect statements. Hallucinations are often attributed to knowledge gaps in LMs, but we hypothesize that in some cases, when justifying previously generated hallucinations, LMs output false claims that they can separately recognize as incorrect. We construct three question-answering datasets where ChatGPT and GPT-4 often state an incorrect answer and offer an explanation with at least one incorrect claim. Crucially, we find that ChatGPT and GPT-4 can identify 67% and 87% of their own mistakes, respectively. We refer to this phenomenon as hallucination snowballing: an LM over-commits to early mistakes, leading to more mistakes that it otherwise would not make.

arXiv comments
Journal reference
DOI
Subjects
cs.CLComputation and Language

On Hacker News

Linked by 4 identifiable commenters in 4 comments across 4 threads, from May 2023 to Dec 2025. Found in 1 HN submission from May 2023 to May 2023. Open submission discussions.

Comments linking this paper, per year.

Discussed in comments4 threads

  1. Project Vend: Phase Two 1 comment

    • We have this write-up on the "soul" and how it was discovered and extracted, straight from the source: https://www.lesswrong.com/posts/vpNG99GhbBoLov9og/claude-4-5... There are many pragmatic reasons to take this "soul data" approach, but we don't know exactly what Anthropic's…

      ACCount37View on HNParent context

  2. An LLM is a lossy encyclopedia 1 comment

    • Assuming you've read OpenAI's paper released this week? https://cdn.openai.com/pdf/d04913be-3f6f-4d2b-b283-ff432ef4a... They attribute these 'compression artefacts' to pre-training, they also reference the original snowballing paper: How Language Model Hallucinations Can…

      jpcompartirView on HNParent context

  3. Hallucination is inevitable: An innate limitation of large language models 1 comment

    • https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=llm+... https://www.medpagetoday.com/ophthalmology/generalophthalmol.... - Survey of Hallucination in Natural Language Generation](https://arxiv.org/abs/2202.03629) (Ji et al., 2022) - [How Language Model Hallucinations Can…

      intendedView on HNParent context

  4. A Logic for Expressing Transformers 1 comment

    • > How do the results of your work inform these predictions that stack-more-layers token-prediction scaling will keep unlocking new abilities? The short answer is that scaling up alone probably won't fix these shortcomings. Better decoding strategies like CoT may help, but the…

      lambdavikingView on HNParent context

Submitted to Hacker News

Connected through shared HN threads, not inferred similarity or a recommendation.