# Analyzing the Source and Target Contributions to Predictions in Neural Machine Translation

arXiv: https://arxiv.org/abs/2010.10907 · HTML: https://mlsanity.com/paper/2010.10907/

## Key facts

- Cited by 71 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 21 Oct 2020.

## Built on by (most cited first, 30 of 71)

- [Explainability for Large Language Models: A Survey](https://mlsanity.com/paper/2309.01029/) (2023)
- [Memorization Without Overfitting: Analyzing the Training Dynamics of Large Language Models](https://mlsanity.com/paper/2205.10770/) (2022)
- [The Curious Case of Hallucinations in Neural Machine Translation](https://mlsanity.com/paper/2104.06683/) (2021)
- [All Bark and No Bite: Rogue Dimensions in Transformer Language Models Obscure Representational Quality](https://mlsanity.com/paper/2109.04404/) (2021)
- [AttnLRP: Attention-Aware Layer-Wise Relevance Propagation for Transformers](https://mlsanity.com/paper/2402.05602/) (2024)
- [Inseq: An Interpretability Toolkit for Sequence Generation Models](https://mlsanity.com/paper/2302.13942/) (2023)
- [Detecting and Mitigating Hallucinations in Machine Translation: Model Internal Workings Alone Do Well, Sentence Similarity Even Better](https://mlsanity.com/paper/2212.08597/) (2022)
- [Mutual Information Alleviates Hallucinations in Abstractive Summarization](https://mlsanity.com/paper/2210.13210/) (2022)
- [Looking for a Needle in a Haystack: A Comprehensive Study of Hallucinations in Neural Machine Translation](https://mlsanity.com/paper/2208.05309/) (2022)
- [Counterfactual Interventions Reveal the Causal Effect of Relative Clause Representations on Agreement Prediction](https://mlsanity.com/paper/2105.06965/) (2021)
- [On Early Detection of Hallucinations in Factual Question Answering](https://mlsanity.com/paper/2312.14183/) (2023)
- [DINOISER: Diffused Conditional Sequence Learning by Manipulating Noises](https://mlsanity.com/paper/2302.10025/) (2023)
- [Usable XAI: 10 Strategies Towards Exploiting Explainability in the LLM Era](https://mlsanity.com/paper/2403.08946/) (2024)
- [Towards Opening the Black Box of Neural Machine Translation: Source and Target Interpretations of the Transformer](https://mlsanity.com/paper/2205.11631/) (2022)
- [Understanding and Detecting Hallucinations in Neural Machine Translation via Model Introspection](https://mlsanity.com/paper/2301.07779/) (2023)
- [What do tokens know about their characters and how do they know it?](https://mlsanity.com/paper/2206.02608/) (2022)
- [Exploring the Role of BERT Token Representations to Explain Sentence Probing Results](https://mlsanity.com/paper/2104.01477/) (2021)
- [Rationales for Sequential Predictions](https://mlsanity.com/paper/2109.06387/) (2021)
- [On Behalf of the Stakeholders: Trends in NLP Model Interpretability in the Era of LLMs](https://mlsanity.com/paper/2407.19200/) (2024)
- [Pre-Trained Multilingual Sequence-to-Sequence Models: A Hope for Low-Resource Language Translation?](https://mlsanity.com/paper/2203.08850/) (2022)
- [On Explaining Your Explanations of BERT: An Empirical Study with Sequence Classification](https://mlsanity.com/paper/2101.00196/) (2021)
- [Language Modeling, Lexical Translation, Reordering: The Training Process of NMT through the Lens of Classical SMT](https://mlsanity.com/paper/2109.01396/) (2021)
- [Reducing Position Bias in Simultaneous Machine Translation with Length-Aware Framework](https://mlsanity.com/paper/2203.09053/) (2022)
- [Are Character-level Translations Worth the Wait? Comparing ByT5 and mT5 for Machine Translation](https://mlsanity.com/paper/2302.14220/) (2023)
- [Optimal Transport for Unsupervised Hallucination Detection in Neural Machine Translation](https://mlsanity.com/paper/2212.09631/) (2022)
- [Attention Weights in Transformer NMT Fail Aligning Words Between Sequences but Largely Explain Model Predictions](https://mlsanity.com/paper/2109.05853/) (2021)
- [Teacher Forcing Recovers Reward Functions for Text Generation](https://mlsanity.com/paper/2210.08708/) (2022)
- [A Close Look at Decomposition-based XAI-Methods for Transformer Language Models](https://mlsanity.com/paper/2502.15886/) (2025)
- [Dissecting Generation Modes for Abstractive Summarization Models via Ablation and Attribution](https://mlsanity.com/paper/2106.01518/) (2021)
- [On Measuring Context Utilization in Document-Level MT Systems](https://mlsanity.com/paper/2402.01404/) (2024)

## Abstract (arXiv)

In Neural Machine Translation (and, more generally, conditional language modeling), the generation of a target token is influenced by two types of context: the source and the prefix of the target sequence. While many attempts to understand the internal workings of NMT models have been made, none of them explicitly evaluates relative source and target contributions to a generation decision. We argue that this relative contribution can be evaluated by adopting a variant of Layerwise Relevance Propagation (LRP). Its underlying 'conservation principle' makes relevance propagation unique: differently from other methods, it evaluates not an abstract quantity reflecting token importance, but the proportion of each token's influence. We extend LRP to the Transformer and conduct an analysis of NMT models which explicitly evaluates the source and target relative contributions to the generation process. We analyze changes in these contributions when conditioning on different types of prefixes, when varying the training objective or the amount of training data, and during the training process. We find that models trained with more data tend to rely on source information more and to have more sharp token contributions; the training process is non-monotonic with several stages of different nature.