# In-context Learning Generalizes, But Not Always Robustly: The Case of Syntax

arXiv: https://arxiv.org/abs/2311.07811 · HTML: https://mlsanity.com/paper/2311.07811/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- Builds on 40 of its 53 references that are in the map.
- First posted on arXiv 13 Nov 2023.

## Builds on (most cited first, 30 of 40)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://mlsanity.com/paper/1910.13461/) (2019)
- [PaLM: Scaling Language Modeling with Pathways](https://mlsanity.com/paper/2204.02311/) (2022)
- [Finetuned Language Models Are Zero-Shot Learners](https://mlsanity.com/paper/2109.01652/) (2021)
- [A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference](https://mlsanity.com/paper/1704.05426/) (2017)
- [Code Llama: Open Foundation Models for Code](https://mlsanity.com/paper/2308.12950/) (2023)
- [Rethinking the Role of Demonstrations: What Makes In-Context Learning Work?](https://mlsanity.com/paper/2202.12837/) (2022)
- [Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting](https://mlsanity.com/paper/2305.04388/) (2023)
- [Right for the Wrong Reasons: Diagnosing Syntactic Heuristics in Natural Language Inference](https://mlsanity.com/paper/1902.01007/) (2019)
- [What learning algorithm is in-context learning? Investigations with linear models](https://mlsanity.com/paper/2211.15661/) (2022)
- [BLiMP: The Benchmark of Linguistic Minimal Pairs for English](https://mlsanity.com/paper/1912.00582/) (2019)
- [MetaICL: Learning to Learn In Context](https://mlsanity.com/paper/2110.15943/) (2021)
- [Larger language models do in-context learning differently](https://mlsanity.com/paper/2303.03846/) (2023)
- [Assessing the Ability of LSTMs to Learn Syntax-Sensitive Dependencies](https://mlsanity.com/paper/1611.01368/) (2016)
- [Language Models Are Greedy Reasoners: A Systematic Formal Analysis of Chain-of-Thought](https://mlsanity.com/paper/2210.01240/) (2022)
- [Do Prompt-Based Models Really Understand the Meaning of their Prompts?](https://mlsanity.com/paper/2109.01247/) (2021)
- [Prompting GPT-3 To Be Reliable](https://mlsanity.com/paper/2210.09150/) (2022)
- [A Comprehensive Capability Analysis of GPT-3 and GPT-3.5 Series Models](https://mlsanity.com/paper/2303.10420/) (2023)
- [Targeted Syntactic Evaluation of Language Models](https://mlsanity.com/paper/1808.09031/) (2018)
- [Neural Theory-of-Mind? On the Limits of Social Intelligence in Large LMs](https://mlsanity.com/paper/2210.13312/) (2022)
- [Inverse Scaling: When Bigger Isn't Better](https://mlsanity.com/paper/2306.09479/) (2023)
- [A Systematic Assessment of Syntactic Generalization in Neural Language Models](https://mlsanity.com/paper/2005.03692/) (2020)
- [Language Models of Code are Few-Shot Commonsense Learners](https://mlsanity.com/paper/2210.07128/) (2022)
- [What In-Context Learning "Learns" In-Context: Disentangling Task Recognition and Task Learning](https://mlsanity.com/paper/2305.09731/) (2023)
- [Sudden Drops in the Loss: Syntax Acquisition, Phase Transitions, and Simplicity Bias in MLMs](https://mlsanity.com/paper/2309.07311/) (2023)

## Built on by (most cited first, 22 of 22)

- [How Does Code Pretraining Affect Language Model Task Performance?](https://mlsanity.com/paper/2409.04556/) (2024)
- [Learning Syntax Without Planting Trees: Understanding Hierarchical Generalization in Transformers](https://mlsanity.com/paper/2404.16367/) (2024)
- [AdaReasoner: Adaptive Reasoning Enables More Flexible Thinking in Large Language Models](https://mlsanity.com/paper/2505.17312/) (2025)
- [Characterizing the Role of Similarity in the Property Inferences of Language Models](https://mlsanity.com/paper/2410.22590/) (2024)
- [Sometimes I am a Tree: Data Drives Unstable Hierarchical Generalization](https://mlsanity.com/paper/2412.04619/) (2024)
- [Optimizing Decomposition for Optimal Claim Verification](https://mlsanity.com/paper/2503.15354/) (2025)
- [Look Further Ahead: Testing the Limits of GPT-4 in Path Planning](https://mlsanity.com/paper/2406.12000/) (2024)
- [Rapid Word Learning Through Meta In-Context Learning](https://mlsanity.com/paper/2502.14791/) (2025)
- [ObscuraCoder: Powering Efficient Code LM Pre-Training Via Obfuscation Grounding](https://mlsanity.com/paper/2504.00019/) (2025)
- [ICL CIPHERS: Quantifying "Learning" in In-Context Learning via Substitution Ciphers](https://mlsanity.com/paper/2504.19395/) (2025)
- [Improving the OOD Performance of Closed-Source LLMs on NLI Through Strategic Data Selection](https://mlsanity.com/paper/2505.20209/) (2025)
- [Experimental Contexts Can Facilitate Robust Semantic Property Inference in Language Models, but Inconsistently](https://mlsanity.com/paper/2401.06640/) (2024)
- [Exploring Explanations Improves the Robustness of In-Context Learning](https://mlsanity.com/paper/2506.02378/) (2025)
- [Simulating Human Memory with Language Models](https://mlsanity.com/paper/2605.25680/) (2026)
- [Learning Linear Regression with Low-Rank Tasks in-Context](https://mlsanity.com/paper/2510.04548/) (2025)
- [Bearing Syntactic Fruit with Stack-Augmented Neural Networks](https://mlsanity.com/paper/2511.03547/) (2025)
- [Analogical Structure, Minimal Contextual Cues and Contrastive Distractors: Input Design for Sample-Efficient Linguistic Rule Induction](https://mlsanity.com/paper/2511.10441/) (2025)
- [Towards Effective In-context Cross-domain Knowledge Transfer via Domain-invariant-neurons-based Retrieval](https://mlsanity.com/paper/2604.05383/) (2026)
- [Reason Analogically via Cross-domain Prior Knowledge: An Empirical Study of Cross-domain Knowledge Transfer for In-Context Learning](https://mlsanity.com/paper/2604.05396/) (2026)
- [Towards Contamination Resistant Benchmarks](https://mlsanity.com/paper/2505.08389/) (2025)
- [The Dog the Cat Chased Stumped the Model: Measuring When Language Models Abandon Structure for Shortcuts](https://mlsanity.com/paper/2510.20543/) (2025)
- [Towards Spec Learning: Inference-Time Alignment from Preference Pairs](https://mlsanity.com/paper/2606.24004/) (2026)

## Abstract (arXiv)

In-context learning (ICL) is now a common method for teaching large language models (LLMs) new tasks: given labeled examples in the input context, the LLM learns to perform the task without weight updates. Do models guided via ICL infer the underlying structure of the task defined by the context, or do they rely on superficial heuristics that only generalize to identically distributed examples? We address this question using transformations tasks and an NLI task that assess sensitivity to syntax - a requirement for robust language understanding. We further investigate whether out-of-distribution generalization can be improved via chain-of-thought prompting, where the model is provided with a sequence of intermediate computation steps that illustrate how the task ought to be performed. In experiments with models from the GPT, PaLM, and Llama 2 families, we find large variance across LMs. The variance is explained more by the composition of the pre-training corpus and supervision methods than by model size; in particular, models pre-trained on code generalize better, and benefit more from chain-of-thought prompting.