# Decomposing Natural Logic Inferences in Neural NLI

arXiv: https://arxiv.org/abs/2112.08289 · HTML: https://mlsanity.com/paper/2112.08289/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 34 references that are in the map.
- First posted on arXiv 15 Dec 2021.

## Builds on (most cited first, 22 of 22)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [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)
- [UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction](https://mlsanity.com/paper/1802.03426/) (2018)
- [A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference](https://mlsanity.com/paper/1704.05426/) (2017)
- [A large annotated corpus for learning natural language inference](https://mlsanity.com/paper/1508.05326/) (2015)
- [Understanding intermediate layers using linear classifier probes](https://mlsanity.com/paper/1610.01644/) (2016)
- [Right for the Wrong Reasons: Diagnosing Syntactic Heuristics in Natural Language Inference](https://mlsanity.com/paper/1902.01007/) (2019)
- [Designing and Interpreting Probes with Control Tasks](https://mlsanity.com/paper/1909.03368/) (2019)
- [Probing Pretrained Language Models for Lexical Semantics](https://mlsanity.com/paper/2010.05731/) (2020)
- [Question and Answer Test-Train Overlap in Open-Domain Question Answering Datasets](https://mlsanity.com/paper/2008.02637/) (2020)
- [Probing Natural Language Inference Models through Semantic Fragments](https://mlsanity.com/paper/1909.07521/) (2019)
- [Pareto Probing: Trading Off Accuracy for Complexity](https://mlsanity.com/paper/2010.02180/) (2020)
- [Can neural networks understand monotonicity reasoning?](https://mlsanity.com/paper/1906.06448/) (2019)
- [Probing Linguistic Systematicity](https://mlsanity.com/paper/2005.04315/) (2020)
- [HELP: A Dataset for Identifying Shortcomings of Neural Models in Monotonicity Reasoning](https://mlsanity.com/paper/1904.12166/) (2019)
- [NeuralLog: Natural Language Inference with Joint Neural and Logical Reasoning](https://mlsanity.com/paper/2105.14167/) (2021)
- [MonaLog: a Lightweight System for Natural Language Inference Based on Monotonicity](https://mlsanity.com/paper/1910.08772/) (2019)
- [MedNLI Is Not Immune: Natural Language Inference Artifacts in the Clinical Domain](https://mlsanity.com/paper/2106.01491/) (2021)
- [LangPro: Natural Language Theorem Prover](https://mlsanity.com/paper/1708.09417/) (2017)
- [Does My Representation Capture X? Probe-Ably](https://mlsanity.com/paper/2104.05807/) (2021)
- [Supporting Context Monotonicity Abstractions in Neural NLI Models](https://mlsanity.com/paper/2105.08008/) (2021)

## Built on by (most cited first, 4 of 4)

- [Interventional Probing in High Dimensions: An NLI Case Study](https://mlsanity.com/paper/2304.10346/) (2023)
- [Estimating the Causal Effects of Natural Logic Features in Neural NLI Models](https://mlsanity.com/paper/2305.08572/) (2023)
- [Synthetic Dataset for Evaluating Complex Compositional Knowledge for Natural Language Inference](https://mlsanity.com/paper/2307.05034/) (2023)
- [Semantic Abstraction for Natural Language Inference: a Methodological Framework for Discovering and Compensating Semantic Knowledge and Reasoning Gaps in Large Language Models](https://mlsanity.com/paper/2609.26610/) (2026)

## Abstract (arXiv)

In the interest of interpreting neural NLI models and their reasoning strategies, we carry out a systematic probing study which investigates whether these models capture the crucial semantic features central to natural logic: monotonicity and concept inclusion. Correctly identifying valid inferences in downward-monotone contexts is a known stumbling block for NLI performance, subsuming linguistic phenomena such as negation scope and generalized quantifiers. To understand this difficulty, we emphasize monotonicity as a property of a context and examine the extent to which models capture monotonicity information in the contextual embeddings which are intermediate to their decision making process. Drawing on the recent advancement of the probing paradigm, we compare the presence of monotonicity features across various models. We find that monotonicity information is notably weak in the representations of popular NLI models which achieve high scores on benchmarks, and observe that previous improvements to these models based on fine-tuning strategies have introduced stronger monotonicity features together with their improved performance on challenge sets.