# Large Language Models are Easily Confused: A Quantitative Metric, Security Implications and Typological Analysis

arXiv: https://arxiv.org/abs/2410.13237 · HTML: https://mlsanity.com/paper/2410.13237/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 34 of its 67 references that are in the map.
- First posted on arXiv 17 Oct 2024.

## Builds on (most cited first, 30 of 34)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [mT5: A massively multilingual pre-trained text-to-text transformer](https://mlsanity.com/paper/2010.11934/) (2020)
- [A Call for Clarity in Reporting BLEU Scores](https://mlsanity.com/paper/1804.08771/) (2018)
- [Bag of Tricks for Efficient Text Classification](https://mlsanity.com/paper/1607.01759/) (2016)
- [How multilingual is Multilingual BERT?](https://mlsanity.com/paper/1906.01502/) (2019)
- [Low-Resource Languages Jailbreak GPT-4](https://mlsanity.com/paper/2310.02446/) (2023)
- [Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation](https://mlsanity.com/paper/2004.11867/) (2020)
- [Information Leakage in Embedding Models](https://mlsanity.com/paper/2004.00053/) (2020)
- [Text Embeddings Reveal (Almost) As Much As Text](https://mlsanity.com/paper/2310.06816/) (2023)
- [Okapi: Instruction-tuned Large Language Models in Multiple Languages with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2307.16039/) (2023)
- [Factuality Challenges in the Era of Large Language Models](https://mlsanity.com/paper/2310.05189/) (2023)
- [Understanding and Mitigating Language Confusion in LLMs](https://mlsanity.com/paper/2406.20052/) (2024)
- [Differentially Private Representation for NLP: Formal Guarantee and An Empirical Study on Privacy and Fairness](https://mlsanity.com/paper/2010.01285/) (2020)
- [Overcoming Catastrophic Forgetting in Zero-Shot Cross-Lingual Generation](https://mlsanity.com/paper/2205.12647/) (2022)
- [Mitigating Hallucinations and Off-target Machine Translation with Source-Contrastive and Language-Contrastive Decoding](https://mlsanity.com/paper/2309.07098/) (2023)
- [Learning Language Representations for Typology Prediction](https://mlsanity.com/paper/1707.09569/) (2017)
- [Text Embedding Inversion Security for Multilingual Language Models](https://mlsanity.com/paper/2401.12192/) (2024)
- [mmT5: Modular Multilingual Pre-Training Solves Source Language Hallucinations](https://mlsanity.com/paper/2305.14224/) (2023)
- [MTG: A Benchmark Suite for Multilingual Text Generation](https://mlsanity.com/paper/2108.07140/) (2021)
- [Why Does Zero-Shot Cross-Lingual Generation Fail? An Explanation and a Solution](https://mlsanity.com/paper/2305.17325/) (2023)
- [Multilingual Blending: LLM Safety Alignment Evaluation with Language Mixture](https://mlsanity.com/paper/2407.07342/) (2024)
- [Continuous multilinguality with language vectors](https://mlsanity.com/paper/1612.07486/) (2016)
- [What do Language Representations Really Represent?](https://mlsanity.com/paper/1901.02646/) (2019)
- [fugashi, a Tool for Tokenizing Japanese in Python](https://mlsanity.com/paper/2010.06858/) (2020)
- [Conceptual similarity and communicative need shape colexification: an experimental study](https://mlsanity.com/paper/2103.11024/) (2021)
- [TuBA: Cross-Lingual Transferability of Backdoor Attacks in LLMs with Instruction Tuning](https://mlsanity.com/paper/2404.19597/) (2024)
- [Key ingredients for effective zero-shot cross-lingual knowledge transfer in generative tasks](https://mlsanity.com/paper/2402.12279/) (2024)
- [Against All Odds: Overcoming Typology, Script, and Language Confusion in Multilingual Embedding Inversion Attacks](https://mlsanity.com/paper/2408.11749/) (2024)

## Built on by (most cited first, 7 of 7)

- [Advances and Challenges in Foundation Agents: From Brain-Inspired Intelligence to Evolutionary, Collaborative, and Safe Systems](https://mlsanity.com/paper/2504.01990/) (2025)
- [Trustworthy Machine Learning via Memorization and the Granular Long-Tail: A Survey on Interactions, Tradeoffs, and Beyond](https://mlsanity.com/paper/2503.07501/) (2025)
- [Shared Path: Unraveling Memorization in Multilingual LLMs through Language Similarities](https://mlsanity.com/paper/2505.15722/) (2025)
- [Controlling Language Confusion in Multilingual LLMs](https://mlsanity.com/paper/2505.19116/) (2025)
- [LAGO: Few-shot Crosslingual Embedding Inversion Attacks via Language Similarity-Aware Graph Optimization](https://mlsanity.com/paper/2505.16008/) (2025)
- [Evaluating Multilingual and Code-Switched Alignment in LLMs via Synthetic Natural Language Inference](https://mlsanity.com/paper/2508.14735/) (2025)
- [Are the LLMs Capable of Maintaining at Least the Language Genus?](https://mlsanity.com/paper/2510.21561/) (2025)

## Abstract (arXiv)

Language Confusion is a phenomenon where Large Language Models (LLMs) generate text that is neither in the desired language, nor in a contextually appropriate language. This phenomenon presents a critical challenge in text generation by LLMs, often appearing as erratic and unpredictable behavior. We hypothesize that there are linguistic regularities to this inherent vulnerability in LLMs and shed light on patterns of language confusion across LLMs. We introduce a novel metric, Language Confusion Entropy, designed to directly measure and quantify this confusion, based on language distributions informed by linguistic typology and lexical variation. Comprehensive comparisons with the Language Confusion Benchmark (Marchisio et al., 2024) confirm the effectiveness of our metric, revealing patterns of language confusion across LLMs. We further link language confusion to LLM security, and find patterns in the case of multilingual embedding inversion attacks. Our analysis demonstrates that linguistic typology offers theoretically grounded interpretation, and valuable insights into leveraging language similarities as a prior for LLM alignment and security.