# The Larger They Are, the Harder They Fail: Language Models do not Recognize Identifier Swaps in Python

arXiv: https://arxiv.org/abs/2305.15507 · HTML: https://mlsanity.com/paper/2305.15507/

## Key facts

- Cited by 29 papers in the map (data through 14 Oct 2026).
- Builds on 28 of its 40 references that are in the map.
- First posted on arXiv 24 May 2023.

## Builds on (most cited first, 28 of 28)

- [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)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [PaLM: Scaling Language Modeling with Pathways](https://mlsanity.com/paper/2204.02311/) (2022)
- [Large Language Models are Zero-Shot Reasoners](https://mlsanity.com/paper/2205.11916/) (2022)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [Scaling Instruction-Finetuned Language Models](https://mlsanity.com/paper/2210.11416/) (2022)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [Neural Machine Translation of Rare Words with Subword Units](https://mlsanity.com/paper/1508.07909/) (2015)
- [The Pile: An 800GB Dataset of Diverse Text for Language Modeling](https://mlsanity.com/paper/2101.00027/) (2020)
- [Survey of Hallucination in Natural Language Generation](https://mlsanity.com/paper/2202.03629/) (2022)
- [Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models](https://mlsanity.com/paper/2206.04615/) (2022)
- [Shortcut Learning in Deep Neural Networks](https://mlsanity.com/paper/2004.07780/) (2020)
- [Competition-Level Code Generation with AlphaCode](https://mlsanity.com/paper/2203.07814/) (2022)
- [Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges](https://mlsanity.com/paper/2104.13478/) (2021)
- [Group Equivariant Convolutional Networks](https://mlsanity.com/paper/1602.07576/) (2016)
- [CodeGen: An Open Large Language Model for Code with Multi-Turn Program Synthesis](https://mlsanity.com/paper/2203.13474/) (2022)
- [BBQ: A Hand-Built Bias Benchmark for Question Answering](https://mlsanity.com/paper/2110.08193/) (2021)
- [True Few-Shot Learning with Language Models](https://mlsanity.com/paper/2105.11447/) (2021)
- [Scaling Laws for Transfer](https://mlsanity.com/paper/2102.01293/) (2021)
- [A parallel corpus of Python functions and documentation strings for automated code documentation and code generation](https://mlsanity.com/paper/1707.02275/) (2017)
- [Learning Python Code Suggestion with a Sparse Pointer Network](https://mlsanity.com/paper/1611.08307/) (2016)
- [BERT is not The Count: Learning to Match Mathematical Statements with Proofs](https://mlsanity.com/paper/2302.09350/) (2023)
- [Equivariant MuZero](https://mlsanity.com/paper/2302.04798/) (2023)

## Built on by (most cited first, 29 of 29)

- [CRUXEval: A Benchmark for Code Reasoning, Understanding and Execution](https://mlsanity.com/paper/2401.03065/) (2024)
- [Reasoning or Reciting? Exploring the Capabilities and Limitations of Language Models Through Counterfactual Tasks](https://mlsanity.com/paper/2307.02477/) (2023)
- [Inverse Scaling: When Bigger Isn't Better](https://mlsanity.com/paper/2306.09479/) (2023)
- [Vulnerability Detection with Code Language Models: How Far Are We?](https://mlsanity.com/paper/2403.18624/) (2024)
- [Exposing Attention Glitches with Flip-Flop Language Modeling](https://mlsanity.com/paper/2306.00946/) (2023)
- [LILO: Learning Interpretable Libraries by Compressing and Documenting Code](https://mlsanity.com/paper/2310.19791/) (2023)
- [CodeMind: Evaluating Large Language Models for Code Reasoning](https://mlsanity.com/paper/2402.09664/) (2024)
- [Large Language Model Reasoning Failures](https://mlsanity.com/paper/2602.06176/) (2026)
- [The Counterfeit Conundrum: Can Code Language Models Grasp the Nuances of Their Incorrect Generations?](https://mlsanity.com/paper/2402.19475/) (2024)
- [Machine-assisted quantitizing designs: augmenting humanities and social sciences with artificial intelligence](https://mlsanity.com/paper/2309.14379/) (2023)
- [Towards Interpretable Sequence Continuation: Analyzing Shared Circuits in Large Language Models](https://mlsanity.com/paper/2311.04131/) (2023)
- [Coffee: Boost Your Code LLMs by Fixing Bugs with Feedback](https://mlsanity.com/paper/2311.07215/) (2023)
- [What can Large Language Models Capture about Code Functional Equivalence?](https://mlsanity.com/paper/2408.11081/) (2024)
- [Interpreting Context Look-ups in Transformers: Investigating Attention-MLP Interactions](https://mlsanity.com/paper/2402.15055/) (2024)
- [An Investigation of LLMs' Inefficacy in Understanding Converse Relations](https://mlsanity.com/paper/2310.05163/) (2023)
- [Teacher-Student Training for Debiasing: General Permutation Debiasing for Large Language Models](https://mlsanity.com/paper/2403.13590/) (2024)
- [Dialogue-based generation of self-driving simulation scenarios using Large Language Models](https://mlsanity.com/paper/2310.17372/) (2023)
- [A Tool for In-depth Analysis of Code Execution Reasoning of Large Language Models](https://mlsanity.com/paper/2501.18482/) (2025)
- [Are Large Language Models Robust in Understanding Code Against Semantics-Preserving Mutations?](https://mlsanity.com/paper/2505.10443/) (2025)
- [Scaling Behavior of Machine Translation with Large Language Models under Prompt Injection Attacks](https://mlsanity.com/paper/2403.09832/) (2024)
- [A Critical Study of What Code-LLMs (Do Not) Learn](https://mlsanity.com/paper/2406.11930/) (2024)
- [Do Current Language Models Support Code Intelligence for R Programming Language?](https://mlsanity.com/paper/2410.07793/) (2024)
- [Guess & Sketch: Language Model Guided Transpilation](https://mlsanity.com/paper/2309.14396/) (2023)
- [Models Can and Should Embrace the Communicative Nature of Human-Generated Math](https://mlsanity.com/paper/2409.17005/) (2024)
- [Program Semantic Inequivalence Game with Large Language Models](https://mlsanity.com/paper/2505.03818/) (2025)
- [Open the Oyster: Empirical Evaluation and Improvement of Code Reasoning Confidence in LLMs](https://mlsanity.com/paper/2511.02197/) (2025)
- [Names Don't Matter: Symbol-Invariant Transformer for Open-Vocabulary Learning](https://mlsanity.com/paper/2601.23169/) (2026)
- [DEVAL: A Framework for Evaluating and Improving the Derivation Capability of Large Language Models](https://mlsanity.com/paper/2511.14813/) (2025)
- [Position: Quantum Program Generation Must Prioritize Validity Over Probabilistic Scaling](https://mlsanity.com/paper/2607.15313/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) have successfully been applied to code generation tasks, raising the question of how well these models understand programming. Typical programming languages have invariances and equivariances in their semantics that human programmers intuitively understand and exploit, such as the (near) invariance to the renaming of identifiers. We show that LLMs not only fail to properly generate correct Python code when default function names are swapped, but some of them even become more confident in their incorrect predictions as the model size increases, an instance of the recently discovered phenomenon of Inverse Scaling, which runs contrary to the commonly observed trend of increasing prediction quality with increasing model size. Our findings indicate that, despite their astonishing typical-case performance, LLMs still lack a deep, abstract understanding of the content they manipulate, making them unsuitable for tasks that statistically deviate from their training data, and that mere scaling is not enough to achieve such capability.