mlsanityData to 7 Oct 2026

Hypothesis-only Biases in Large Language Model-Elicited Natural Language Inference

arXiv 2410.08996 ↗First posted 11 Oct 2024cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

We test whether replacing crowdsource workers with LLMs to write Natural Language Inference (NLI) hypotheses similarly results in annotation artifacts. We recreate a portion of the Stanford NLI corpus using GPT-4, Llama-2 and Mistral 7b, and train hypothesis-only classifiers to determine whether LLM-elicited hypotheses contain annotation artifacts. On our LLM-elicited NLI datasets, BERT-based hypothesis-only classifiers achieve between 86-96% accuracy, indicating these datasets contain hypothesis-only artifacts. We also find frequent "give-aways" in LLM-generated hypotheses, e.g. the phrase "swimming in a pool" appears in more than 10,000 contradictions generated by GPT-4. Our analysis provides empirical evidence that well-attested biases in NLI can persist in LLM-generated data.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 25 of its 42 references that are in the map. First posted on arXiv 11 Oct 2024.

Builds on

25 of 42 references.

References · 25

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 12,865
  3. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,674
  4. 2022BLOOM: A 176B-Parameter Open-Access Multilingual Language Modelcited by 1,968
  5. 2015A large annotated corpus for learning natural language inferencecited by 1,837
  6. 2019SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systemscited by 1,718
  7. 2022In-context Learning and Induction Headscited by 930
  8. 2023ChatGPT Outperforms Crowd-Workers for Text-Annotation Taskscited by 849
  9. 2018Annotation Artifacts in Natural Language Inference Datacited by 656
  10. 2024OLMo: Accelerating the Science of Language Modelscited by 599
  11. 2023The Falcon Series of Open Language Modelscited by 526
  12. 2023Benchmarking Large Language Models for News Summarizationcited by 444
13 more
  1. 2022Unnatural Instructions: Tuning Language Models with (Almost) No Human Laborcited by 395
  2. 2018Hypothesis Only Baselines in Natural Language Inferencecited by 291
  3. 2021Polyjuice: Generating Counterfactuals for Explaining, Evaluating, and Improving Modelscited by 219
  4. 2023AnnoLLM: Making Large Language Models to Be Better Crowdsourced Annotatorscited by 171
  5. 2021Reframing Human-AI Collaboration for Generating Free-Text Explanationscited by 125
  6. 2023Increasing Diversity While Maintaining Accuracy: Text Data Generation with Large Language Models and Human Interventionscited by 121
  7. 2023Artificial Artificial Artificial Intelligence: Crowd Workers Widely Use Large Language Models for Text Production Taskscited by 102
  8. 2018How Much Reading Does Reading Comprehension Require? A Critical Investigation of Popular Benchmarkscited by 86
  9. 2022Rethinking the Role of Scale for In-Context Learning: An Interpretability-based Case Study at 66 Billion Scalecited by 64
  10. 2018Performance Impact Caused by Hidden Bias of Training Data for Recognizing Textual Entailmentcited by 58
  11. 2017The Effect of Different Writing Tasks on Linguistic Style: A Case Study of the ROC Story Cloze Taskcited by 39
  12. 2023Large Language Models as Annotators: Enhancing Generalization of NLP Models at Minimal Costcited by 33
  13. 2024Genie: Achieving Human Parity in Content-Grounded Datasets Generationcited by 31

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Biases in Large Language Model-Elicited Text: A Case Study in Natural Language Inference
  2. 2025Neutralizing Bias in LLM Reasoning using Entailment Graphs
  3. 2023No Strong Feelings One Way or Another: Re-operationalizing Neutrality in Natural Language Inference
  4. 2024A Peek into Token Bias: Large Language Models Are Not Yet Genuine Reasoners
  5. 2024Provenance: A Light-weight Fact-checker for Retrieval Augmented LLM Generation Output
  6. 2026Biases in the Blind Spot: Detecting What LLMs Fail to Mention
  7. 2023LLMs as Factual Reasoners: Insights from Existing Benchmarks and Beyond
  8. 2024Long-form factuality in large language models