# What Can We Learn from Collective Human Opinions on Natural Language Inference Data?

arXiv: https://arxiv.org/abs/2010.03532 · HTML: https://mlsanity.com/paper/2010.03532/

## Key facts

- Cited by 153 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 7 Oct 2020.

## Built on by (most cited first, 30 of 153)

- [MiniCheck: Efficient Fact-Checking of LLMs on Grounding Documents](https://mlsanity.com/paper/2404.10774/) (2024)
- [The 'Problem' of Human Label Variation: On Ground Truth in Data, Modeling and Evaluation](https://mlsanity.com/paper/2211.02570/) (2022)
- [Teach Me to Explain: A Review of Datasets for Explainable Natural Language Processing](https://mlsanity.com/paper/2102.12060/) (2021)
- [We're Afraid Language Models Aren't Modeling Ambiguity](https://mlsanity.com/paper/2304.14399/) (2023)
- [Modeling Future Conversation Turns to Teach LLMs to Ask Clarifying Questions](https://mlsanity.com/paper/2410.13788/) (2024)
- [Stop Measuring Calibration When Humans Disagree](https://mlsanity.com/paper/2210.16133/) (2022)
- [Metacognitive Prompting Improves Understanding in Large Language Models](https://mlsanity.com/paper/2308.05342/) (2023)
- [Bridging the Gap: A Survey on Integrating (Human) Feedback for Natural Language Generation](https://mlsanity.com/paper/2305.00955/) (2023)
- [The Past, Present and Better Future of Feedback Learning in Large Language Models for Subjective Human Preferences and Values](https://mlsanity.com/paper/2310.07629/) (2023)
- [SimBench: Benchmarking the Ability of Large Language Models to Simulate Human Behaviors](https://mlsanity.com/paper/2510.17516/) (2025)
- [VariErr NLI: Separating Annotation Error from Human Label Variation](https://mlsanity.com/paper/2403.01931/) (2024)
- [You Are What You Annotate: Towards Better Models through Annotator Representations](https://mlsanity.com/paper/2305.14663/) (2023)
- [AmbiFC: Fact-Checking Ambiguous Claims with Evidence](https://mlsanity.com/paper/2104.00640/) (2021)
- [The Perspectivist Paradigm Shift: Assumptions and Challenges of Capturing Human Labels](https://mlsanity.com/paper/2405.05860/) (2024)
- [Natural Language Deduction through Search over Statement Compositions](https://mlsanity.com/paper/2201.06028/) (2022)
- [Are Larger Pretrained Language Models Uniformly Better? Comparing Performance at the Instance Level](https://mlsanity.com/paper/2105.06020/) (2021)
- [Can Large Language Models Capture Dissenting Human Voices?](https://mlsanity.com/paper/2305.13788/) (2023)
- [Distributed NLI: Learning to Predict Human Opinion Distributions for Language Reasoning](https://mlsanity.com/paper/2104.08676/) (2021)
- [Investigating Reasons for Disagreement in Natural Language Inference](https://mlsanity.com/paper/2209.03392/) (2022)
- [The Effectiveness of LLMs as Annotators: A Comparative Overview and Empirical Analysis of Direct Representation](https://mlsanity.com/paper/2405.01299/) (2024)
- [Learning with Different Amounts of Annotation: From Zero to Many Labels](https://mlsanity.com/paper/2109.04408/) (2021)
- [LLMs as Workers in Human-Computational Algorithms? Replicating Crowdsourcing Pipelines with LLMs](https://mlsanity.com/paper/2307.10168/) (2023)
- [Learning to Refine with Fine-Grained Natural Language Feedback](https://mlsanity.com/paper/2407.02397/) (2024)
- [It Takes Two to Tango: Navigating Conceptualizations of NLP Tasks and Measurements of Performance](https://mlsanity.com/paper/2305.09022/) (2023)
- [Nine Judges, Two Effective Votes: Correlated Errors Undermine LLM Evaluation Panels](https://mlsanity.com/paper/2605.29800/) (2026)
- [Analyzing Dynamic Adversarial Training Data in the Limit](https://mlsanity.com/paper/2110.08514/) (2021)
- [SemEval-2024 Shared Task 6: SHROOM, a Shared-task on Hallucinations and Related Observable Overgeneration Mistakes](https://mlsanity.com/paper/2403.07726/) (2024)
- [Fundamental Problems With Model Editing: How Should Rational Belief Revision Work in LLMs?](https://mlsanity.com/paper/2406.19354/) (2024)
- [What Ingredients Make for an Effective Crowdsourcing Protocol for Difficult NLU Data Collection Tasks?](https://mlsanity.com/paper/2106.00794/) (2021)
- [Embracing Ambiguity: Shifting the Training Target of NLI Models](https://mlsanity.com/paper/2106.03020/) (2021)

## Abstract (arXiv)

Despite the subjective nature of many NLP tasks, most NLU evaluations have focused on using the majority label with presumably high agreement as the ground truth. Less attention has been paid to the distribution of human opinions. We collect ChaosNLI, a dataset with a total of 464,500 annotations to study Collective HumAn OpinionS in oft-used NLI evaluation sets. This dataset is created by collecting 100 annotations per example for 3,113 examples in SNLI and MNLI and 1,532 examples in Abductive-NLI. Analysis reveals that: (1) high human disagreement exists in a noticeable amount of examples in these datasets; (2) the state-of-the-art models lack the ability to recover the distribution over human labels; (3) models achieve near-perfect accuracy on the subset of data with a high level of human agreement, whereas they can barely beat a random guess on the data with low levels of human agreement, which compose most of the common errors made by state-of-the-art models on the evaluation sets. This questions the validity of improving model performance on old metrics for the low-agreement part of evaluation datasets. Hence, we argue for a detailed examination of human agreement in future data collection efforts, and evaluating model outputs against the distribution over collective human opinions. The ChaosNLI dataset and experimental scripts are available at https://github.com/easonnie/ChaosNLI