# Can You Learn Semantics Through Next-Word Prediction? The Case of Entailment

arXiv: https://arxiv.org/abs/2402.13956 · HTML: https://mlsanity.com/paper/2402.13956/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 34 references that are in the map.
- First posted on arXiv 21 Feb 2024.

## Builds on (most cited first, 18 of 18)

- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding](https://mlsanity.com/paper/1804.07461/) (2018)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference](https://mlsanity.com/paper/1704.05426/) (2017)
- [The Pile: An 800GB Dataset of Diverse Text for Language Modeling](https://mlsanity.com/paper/2101.00027/) (2020)
- [Character-level Convolutional Networks for Text Classification](https://mlsanity.com/paper/1509.01626/) (2015)
- [Adversarial NLI: A New Benchmark for Natural Language Understanding](https://mlsanity.com/paper/1910.14599/) (2019)
- [Annotation Artifacts in Natural Language Inference Data](https://mlsanity.com/paper/1803.02324/) (2018)
- [Multi-News: a Large-Scale Multi-Document Summarization Dataset and Abstractive Hierarchical Model](https://mlsanity.com/paper/1906.01749/) (2019)
- [Think before you speak: Training Language Models With Pause Tokens](https://mlsanity.com/paper/2310.02226/) (2023)
- [WANLI: Worker and AI Collaboration for Natural Language Inference Dataset Creation](https://mlsanity.com/paper/2201.05955/) (2022)
- [Sudden Drops in the Loss: Syntax Acquisition, Phase Transitions, and Simplicity Bias in MLMs](https://mlsanity.com/paper/2309.07311/) (2023)
- [Provable Limitations of Acquiring Meaning from Ungrounded Form: What Will Future Language Models Understand?](https://mlsanity.com/paper/2104.10809/) (2021)
- [When redundancy is useful: A Bayesian approach to 'overinformative' referring expressions](https://mlsanity.com/paper/1903.08237/) (2019)
- [Entailment Semantics Can Be Extracted from an Ideal Language Model](https://mlsanity.com/paper/2209.12407/) (2022)
- [What company do words keep? Revisiting the distributional semantics of J.R. Firth & Zellig Harris](https://mlsanity.com/paper/2205.07750/) (2022)
- [Transparency Helps Reveal When Language Models Learn Meaning](https://mlsanity.com/paper/2210.07468/) (2022)

## Built on by (most cited first, 9 of 9)

- [Beyond Preferences in AI Alignment](https://mlsanity.com/paper/2408.16984/) (2024)
- [xLSTM-Mixer: Multivariate Time Series Forecasting by Mixing via Scalar Memories](https://mlsanity.com/paper/2410.16928/) (2024)
- [EMTeC: A Corpus of Eye Movements on Machine-Generated Texts](https://mlsanity.com/paper/2408.04289/) (2024)
- [Disentangling Singlish Discourse Particles with Task-Driven Representation](https://mlsanity.com/paper/2409.20366/) (2024)
- [Probing Large Language Models for Scalar Adjective Lexical Semantics and Scalar Diversity Pragmatics](https://mlsanity.com/paper/2404.03301/) (2024)
- [ExpliCa: Evaluating Explicit Causal Reasoning in Large Language Models](https://mlsanity.com/paper/2502.15487/) (2025)
- [A Comparative Investigation of Compositional Syntax and Semantics in DALL-E 2](https://mlsanity.com/paper/2403.12294/) (2024)
- [A Rose by Any Other Name Would Smell as Sweet: Categorical Homotopy Theory for Large Language Models](https://mlsanity.com/paper/2508.10018/) (2025)
- [Learning Concepts, Not Tokens: Self-Supervised Semantic Alignment for Language Models](https://mlsanity.com/paper/2603.29123/) (2026)

## Abstract (arXiv)

Do LMs infer the semantics of text from co-occurrence patterns in their training data? Merrill et al. (2022) argue that, in theory, sentence co-occurrence probabilities predicted by an optimal LM should reflect the entailment relationship of the constituent sentences, but it is unclear whether probabilities predicted by neural LMs encode entailment in this way because of strong assumptions made by Merrill et al. (namely, that humans always avoid redundancy). In this work, we investigate whether their theory can be used to decode entailment relations from neural LMs. We find that a test similar to theirs can decode entailment relations between natural sentences, well above random chance, though not perfectly, across many datasets and LMs. This suggests LMs implicitly model aspects of semantics to predict semantic effects on sentence co-occurrence patterns. However, we find the test that predicts entailment in practice works in the opposite direction to the theoretical test. We thus revisit the assumptions underlying the original test, finding its derivation did not adequately account for redundancy in human-written text. We argue that better accounting for redundancy related to explanations might derive the observed flipped test and, more generally, improve computational models of speakers in linguistics.