# Source-Free MT Evaluation Is Not MT Evaluation

arXiv: https://arxiv.org/abs/2608.20925 · HTML: https://mlsanity.com/paper/2608.20925/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 18 of its 32 references that are in the map.
- First posted on arXiv 21 Aug 2026.

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

- [BERTScore: Evaluating Text Generation with BERT](https://mlsanity.com/paper/1904.09675/) (2019)
- [BLEURT: Learning Robust Metrics for Text Generation](https://mlsanity.com/paper/2004.04696/) (2020)
- [COMET: A Neural Framework for MT Evaluation](https://mlsanity.com/paper/2009.09025/) (2020)
- [Large Language Models Are State-of-the-Art Evaluators of Translation Quality](https://mlsanity.com/paper/2302.14520/) (2023)
- [xCOMET: Transparent Machine Translation Evaluation through Fine-grained Error Detection](https://mlsanity.com/paper/2310.10482/) (2023)
- [GEMBA-MQM: Detecting Translation Quality Error Spans with GPT-4](https://mlsanity.com/paper/2310.13988/) (2023)
- [Gender Bias in Machine Translation](https://mlsanity.com/paper/2104.06001/) (2021)
- [BLEU might be Guilty but References are not Innocent](https://mlsanity.com/paper/2004.06063/) (2020)
- [Getting Gender Right in Neural Machine Translation](https://mlsanity.com/paper/1909.05088/) (2019)
- [The Devil is in the Errors: Leveraging Large Language Models for Fine-grained Machine Translation Evaluation](https://mlsanity.com/paper/2308.07286/) (2023)
- [MetricX-24: The Google Submission to the WMT 2024 Metrics Shared Task](https://mlsanity.com/paper/2410.03983/) (2024)
- [Error Analysis Prompting Enables Human-Like Translation Evaluation in Large Language Models](https://mlsanity.com/paper/2303.13809/) (2023)
- [Machine Translation Meta Evaluation through Translation Accuracy Challenge Sets](https://mlsanity.com/paper/2401.16313/) (2024)
- [Lost in the Source Language: How Large Language Models Evaluate the Quality of Machine Translation](https://mlsanity.com/paper/2401.06568/) (2024)
- [The Inside Story: Towards Better Understanding of Machine Translation Neural Evaluation Metrics](https://mlsanity.com/paper/2305.11806/) (2023)
- [M-MAD: Multidimensional Multi-Agent Debate for Advanced Machine Translation Evaluation](https://mlsanity.com/paper/2412.20127/) (2024)
- [Feeding Two Birds or Favoring One? Adequacy-Fluency Tradeoffs in Evaluation and Meta-Evaluation of Machine Translation](https://mlsanity.com/paper/2509.20287/) (2025)
- [Gender-Neutral Machine Translation Strategies in Practice](https://mlsanity.com/paper/2506.15676/) (2025)

## Abstract (arXiv)

Reference-based metrics remain the standard choice in machine translation evaluation, partly because quality estimation methods often correlate less well with human judgments. As a result, source-free, reference-based evaluation has become the practical norm, even though it is unfaithful to the definition of translation adequacy and unfair to systems whose outputs preserve the source meaning while differing from the reference. This paper argues that adequacy must be judged with respect to the source. A reference is only one possible rendering of the source and may introduce bias, under-specification, or errors. We further argue that source-reference-hypothesis evaluation is fair only when the judge treats the reference as auxiliary evidence rather than as the primary standard. Otherwise, even source-aware evaluation can reduce adequacy to preference towards reference. We show the existing hybrid metrics are highly reliant on reference compared to source. Our argument is not that all automatic MT metrics fail to use the source. Rather, we argue that any evaluation protocol that removes the source, or allows the reference to dominate the source, is structurally incomplete for adequacy evaluation. However, existing MT papers generally prefer reference-based metrics and use QE metrics only when reference is unavailable. We therefore call for QE to be reframed as a primary approach to source-grounded adequacy evaluation, rather than as a fallback motivated by missing references. We further call for hybrid metrics whose designs explicitly prioritize source--hypothesis faithfulness while using references only as complementary evidence.