Pith. sign in

REVIEW 4 major objections 6 minor 50 references

Compensating for Data with Reasoning: Low-Resource Machine Translation with LLMs

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Fragment-Shot Prompting, which retrieves translation examples for each syntactic fragment of the input, substantially improves LLM translation into and between low-resource Ladin variants, and syntactic coverage tracks translation quality.

desk verdict A practical, well-documented prompting recipe for low-resource MT shows big BLEU gains, but the fragment-selection mechanism isn't cleanly separated from prompt-size effects. read the letter →

arxiv 2505.22293 v1 pith:AOQBCISK submitted 2025-05-28 cs.CL

Samuel Frontull , Thomas Ströhle This is my paper
classification cs.CL
keywords Fragment-ShotPromptingin-contextlearninglow-resourcemachinetranslationLadinretrieval-augmentedgenerationsyntacticcoveragepivotedreasoningmodels
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Fragment-Shot Prompting is an in-context learning method that splits the sentence to be translated into contiguous word fragments and, for each fragment, retrieves up to six parallel sentence pairs containing that exact fragment on the source side. The paper claims this syntactic-coverage retrieval substantially improves LLM translation into and between two low-resource Ladin variants, with the Italian-to-Val Badia BLEU score (a standard translation-quality metric) rising from 4.91 under zero-shot prompting to 14.22 with Fragment-Shot for DeepSeek-R1. A pivoted extension chains fragment retrieval through Italian as a pivot, enabling translation between the two Ladin variants when no direct parallel data exists. The paper also claims that models with stronger reasoning abilities make better use of the retrieved examples, and that prompt engineering offers little benefit when translating from Ladin into Italian, where zero-shot prompting already works well. If these claims are right, a small parallel corpus plus a reasoning-capable LLM can substitute for large amounts of training data in low-resource translation.

What carries the argument

The central mechanism is the Fragment-Shot prompt: the input sentence is segmented into contiguous, non-overlapping word sequences, each sequence is looked up as an exact source-side match in the retrieval corpus, and up to six matched sentence pairs are added to the prompt as examples. A sliding window starts at seven words and shrinks to single words until a match is found, and fragments are chosen to avoid overlap so the input is covered completely but not redundantly. Pivoted Fragment-Shot nests this retrieval twice through Italian, first from the source variant to Italian and then from Italian to the target variant, and deliberately excludes exact pivot-sentence matches to force the model to reason across fragments. The coverage statistic — the share of input words for which examples were found — is the quantity the paper correlates with BLEU to argue that syntactic coverage, rather than mere example presence, drives quality.

What would settle it

Build Random-Shot prompts that match Fragment-Shot prompts in number of example pairs and total character count (Table 2 shows Fragment-Shot averages 8,974 characters vs 2,232 for Random-Shot) on the same 175 test sentences and models; if BLEU scores equalize, the fragment-retrieval mechanism is not what drives the gains.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that fragment-level retrieval, not sentence-level similarity, is what makes in-context examples effective for low-resource translation. Fragment-Shot partitions the input into non-overlapping fragments from seven words down to single words, matches each fragment exactly on the source side of a small parallel corpus, and inserts the matched sentence pairs into the prompt before the input. Across GPT-3.5, GPT-4o, o1-mini, Llama-3.3, and DeepSeek-R1, this method gives the highest BLEU scores for Italian-to-Ladin and for translation between the Val Badia and Gherdëina variants, and the fraction of input words covered by retrieved examples correlates with BLEU in exactly those directions. Pivoted Fragment-Shot applies the same retrieval twice through Italian and significantly improves variant-to-variant translation for reasoning-capable models, though it stays below direct Fragment-Shot. The paper concludes that multi-hop reasoning can, at least in part, compensate for missing parallel data.

Load-bearing premise

The central assumption is that Fragment-Shot's gains over Random-Shot come from retrieving examples by syntactic fragment coverage, not from the much longer prompts and many more examples that Fragment-Shot packs into the context.

Editorial extensions

If this is right

  • For Italian-to-Ladin translation, Fragment-Shot raises BLEU over zero-shot for every tested model; the largest gain in the paper is DeepSeek-R1 on Italian-to-Val Badia, from 4.91 to 14.22.
  • Pivoted Fragment-Shot offers a route to translate between language varieties that lack direct parallel data, as long as a pivot language and two bilingual corpora are available, and it works significantly only for reasoning-capable models.
  • Syntactic coverage can serve as an example-selection signal when semantic embedding models are unavailable or unreliable for a language.
  • In the low-resource-to-high-resource direction, elaborate prompting adds little: zero-shot prompting is already competitive, so effort there is better spent elsewhere.
  • Even the best Fragment-Shot LLM output (14.22 BLEU) remains below a specialised NMT model (16.77) for Italian-to-Val Badia, but LLMs outperform that NMT model in the reverse direction, suggesting the two approaches are complementary.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A direct test the paper does not run would be to pad Random-Shot prompts to match Fragment-Shot prompts in character count (Table 2 reports 8,974 vs 2,232 characters on average); if BLEU gains vanish, the specific contribution of fragment retrieval is not yet established.
  • Exact surface-form fragment matching will likely transfer poorly to morphologically rich languages, where one word can appear in many inflected forms; lemmatised or stem-based matching is a natural extension.
  • The reported mean PF prompt size of 24,852 characters implies a practical ceiling: for longer inputs, fragment alignment between source and pivot would be needed to avoid exhausting the model's context window.
  • The coverage–BLEU correlation suggests an adaptive retrieval rule: stop adding examples once syntactic coverage saturates, which would cut the roughly 34-second DeepSeek-R1 inference time reported for Fragment-Shot while preserving most of the quality gain.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper introduces Fragment-Shot Prompting (FS), an in-context learning method for low-resource machine translation that segments the input sentence into contiguous fragments, retrieves up to six parallel sentence pairs per fragment from a small bilingual corpus, and assembles them into a single prompt. It also introduces Pivoted Fragment-Shot (PF), which performs nested retrieval through a pivot language when no direct parallel corpus is available. The authors evaluate FS, PF, zero-shot, and random-shot prompting with five LLMs (GPT-3.5, GPT-4o, o1-mini, Llama-3.3, DeepSeek-R1) on translation among Italian and two Ladin variants, reporting BLEU scores, prompt statistics, a coverage–BLEU correlation analysis, and qualitative examples. They release code and retrieval corpora. The main empirical findings are large BLEU gains for FS over random-shot when translating from Italian into Ladin and between the two Ladin variants, with reasoning-oriented models (especially DeepSeek-R1) benefiting most; PF also improves over baselines for reasoning models in the variant-to-variant direction.

Significance. If the proposed mechanism is validated, the work offers a practical recipe for leveraging small parallel corpora in low-resource MT with LLMs, and PF is an interesting extension for language pairs without direct parallel data. The paper is one of the first to systematically study LLM prompting for Ladin, and the public release of code and corpora is a useful community resource. The finding that reasoning models use retrieved fragments more effectively is intriguing and could guide future work on retrieval-augmented prompting. The main weakness is that the headline FS-vs-RS comparison is confounded by prompt size and example count, and the coverage–BLEU correlation does not control for these factors, so the specific contribution of fragment-based retrieval is not yet established. Given the magnitude of the reported gains, the method is still promising, but the mechanistic claim needs additional controlled experiments.

major comments (4)
  1. [Section 3, Table 2] The comparison between FS and RS conflates retrieval strategy with prompt size and number of in-context examples. Table 2 reports average prompt lengths of 2,232 characters for RS versus 8,974 for FS, and the FS method includes up to six sentence pairs per retrieved fragment, so FS prompts contain many more examples than the fixed 16 in RS. The large BLEU improvements in Table 1 (e.g., Italian→Val Badia, DeepSeek-R1: 6.91 with RS to 14.22 with FS) could therefore be driven largely by example count, prompt length, or increased lexical overlap with the input, rather than by fragment-level segmentation and retrieval. Please add a control that equates prompt size/example count, such as random-shot with the same number of examples as the FS prompt, or a retrieval baseline that selects whole sentences by lexical overlap without fragment decomposition.
  2. [Table 3] The coverage–BLEU correlations are computed only within FS prompts and do not partial out prompt length, number of retrieved examples, or source sentence length. Since coverage is mechanically related to the number of fragments with retrieved examples, the significant positive correlations (0.28–0.53) may reflect a 'more retrieved data' effect rather than syntactic coverage per se. In addition, with 30 correlations tested (5 models × 6 directions) and no multiple-comparison correction, the starred entries should be interpreted cautiously. Please report partial correlations controlling for prompt size and example count, or otherwise separate the coverage effect from the volume-of-examples effect.
  3. [Section 6, Limitations] The test set consists of only 175 sentences, as the authors acknowledge in the Limitations section. The per-system confidence intervals in Table 1 are wide, and the paper does not report confidence intervals on the differences between methods. Claims such as 'FS consistently achieves the highest performance' are stronger than the evidence supports. Please provide bootstrap intervals for the pairwise differences or evaluate on a larger subset of FLORES+ to confirm that the FS gains are robust.
  4. [Section 6] The description of the significance testing is under-specified. The paper states that 'pairwise statistical significance tests using sacrebleu' were performed, but it does not specify the test statistic, the number of resamples, or whether any correction for multiple comparisons was applied across the many pairwise tests. Given the number of comparisons in Table 1, uncorrected p-values may produce spurious underlines and bold entries. Please document the procedure, including the resampling scheme and the multiple-comparison correction.
minor comments (6)
  1. [Abstract] The claim that 'syntactic coverage positively correlating with translation quality' is too broad; Table 3 shows that the correlation is near zero for the Ladin→Italian directions (e.g., 0.04–0.10 for GPT-3.5). Please qualify the claim to the directions into and between low-resource languages.
  2. [Section 3, Pivoted Fragment-Shot] The sentence 'We reduced to 2 for that exceeded the context size of the model' is unclear; please rephrase to specify that the number of retrieved example pairs was reduced to two for sentences that exceeded the model's context window.
  3. [Section 4] 'intution' should be 'intuition'; also, the BLEU scores for untranslated text are an interesting baseline but should be computed with the same tokenizer as the main results.
  4. [Table 3] The header 'pearson correlation' should be capitalized, and the column label 'FS-coverage and pearson correlationFS-coverage–BLEUstatistics' is missing spaces; also, the table does not report the p-values or adjusted p-values for the starred correlations.
  5. [Section 5] The parameter counts for GPT-3.5 and GPT-4o (175B and 200B) are not official and should be cited or removed.
  6. [Section 3, FS method] The description of fragment selection is somewhat underspecified (e.g., how overlapping matches are resolved); the released code helps, but adding a precise algorithmic description or pseudocode in the paper would improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Fragment-Shot retrieval is algorithmic, the coverage–BLEU correlation is empirical, and the only author-overlapping citation is background context rather than load-bearing evidence.

full rationale

The paper introduces Fragment-Shot (FS) and Pivoted Fragment-Shot (PF), both defined as deterministic retrieval procedures over fixed external corpora, with hand-set parameters (sliding window of seven words, up to six examples per fragment, reduced to two or three in PF) that are not fit to the test data. The central claim that FS improves BLEU and that syntactic coverage correlates with translation quality is supported by direct experiments in Tables 1 and 3: coverage is computed by counting source words for which retrieved examples exist, and the Pearson correlations are post-hoc descriptive statistics over the FS setting, not quantities used to generate the BLEU scores. No fitted parameter is renamed as a prediction, and no result reduces by definition to its own input. The only citation overlapping with an author, Frontull and Moser (2024), is used to supply previously published NMT baseline scores and prior back-translation context for Ladin; it is not invoked to justify the fragment-retrieval mechanism or to exclude alternative methods. The Table 2 observation that FS prompts are roughly four times longer than RS prompts (8,974 vs. 2,232 characters) is a legitimate experimental confound for the FS-vs-RS comparison, but it is a correctness or design concern, not circularity: the retrieval method is independent of the measured outcomes. Therefore the derivation chain is self-contained, and the correct circularity finding is no significant circularity.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the assumption that fragment-level lexical retrieval, rather than just prompt size, drives the improvements. The hand-set hyperparameters (window size, example caps, pivot exclusion) are not fitted to test data, so the circularity burden is low, but they are manual choices that affect the magnitude of reported gains.

free parameters (5)
  • Initial sliding window size = 7 words (reduced to 1)
    The starting fragment length affects which examples are retrieved and is a manual choice, not fit to test data.
  • Maximum examples per fragment = 6 for FS, 3 for PF, reduced to 2 when context exceeded
    The cap on retrieved pairs per fragment controls prompt size and potential gains, and is hand-selected.
  • Exact pivot-sentence exclusion = True
    PF deliberately excludes exact Italian sentence matches to force multi-hop reasoning, which shapes the reported PF results.
  • Non-translatable word exclusions = Proper names and assumed non-translatable words
    Coverage calculations subtract manually assumed non-translatable words, affecting the reported coverage-BLEU correlation.
  • Random example selection = Unseeded random
    When several corpus sentences contain a fragment, examples are randomly selected; without a fixed seed this injects variance.
assumptions (5)
  • standard math BLEU is a valid evaluation metric for low-resource MT quality
    The paper uses sacrebleu BLEU as the sole automatic metric without human evaluation or chrF/COMET.
  • standard math sacrebleu bootstrap significance tests are correctly applied
    The paper reports pairwise significance tests using sacrebleu but does not specify resampling details or multiple-testing corrections.
  • domain assumption The dictionary-derived corpora contain accurate parallel sentences
    The Gherdëina-Italian corpus is extracted from a dictionary and the Val Badia-Italian corpus is from a prior publication; errors would propagate into retrieved examples.
  • domain assumption The 175 FLORES+ dev translations by native speakers are correct references
    The translations follow OLDI guidelines, but the subset selection is not described and the test set is withheld.
  • domain assumption Italian is an effective pivot language for Ladin variants
    The closeness of Italian and Ladin is acknowledged as a possible advantage; this may not generalize to distant languages.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Compensating for Data with Reasoning: Low-Resource Machine Translation with LLMs." pith.science (2026). https://pith.science/paper/AOQBCISK

@misc{pith2026250522293,
  author       = {Pith},
  title        = {Pith review of: Compensating for Data with Reasoning: Low-Resource Machine Translation with LLMs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AOQBCISK}},
  note         = {Machine review of arXiv:2505.22293}
}
read the original abstract

Large Language Models (LLMs) have demonstrated strong capabilities in multilingual machine translation, sometimes even outperforming traditional neural systems. However, previous research has highlighted the challenges of using LLMs, particularly with prompt engineering, for low-resource languages. In this work, we introduce Fragment-Shot Prompting, a novel in-context learning method that segments input and retrieves translation examples based on syntactic coverage, along with Pivoted Fragment-Shot, an extension that enables translation without direct parallel data. We evaluate these methods using GPT-3.5, GPT-4o, o1-mini, LLaMA-3.3, and DeepSeek-R1 for translation between Italian and two Ladin variants, revealing three key findings: (1) Fragment-Shot Prompting is effective for translating into and between the studied low-resource languages, with syntactic coverage positively correlating with translation quality; (2) Models with stronger reasoning abilities make more effective use of retrieved knowledge, generally produce better translations, and enable Pivoted Fragment-Shot to significantly improve translation quality between the Ladin variants; and (3) prompt engineering offers limited, if any, improvements when translating from a low-resource to a high-resource language, where zero-shot prompting already yields satisfactory results. We publicly release our code and the retrieval corpora.

Figures

Figures reproduced from arXiv: 2505.22293 by the authors.

Figure 1
Figure 1. Fragment-Shot Prompting the generalization capabilities of LLMs through In-Context Learning (ICL) (Rubin et al., 2022; Cahyawijaya et al., 2024; Dong et al., 2024), using a single RAG-augmented prompt. Specifically, it explores what can be achieved with a small set of parallel sentences available for as retrieval corpus. Our key contributions: (i) We introduce the Frag￾ment-Shot prompting technique, a novel promptin… view at source ↗
Figure 2
Figure 2. Example of Pivoted-Fragments Prompting and the corresponding reasoning employed by different LLMs. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

50 extracted references · 8 canonical work pages

  1. [1]

    Sweta Agrawal, Chunting Zhou, Mike Lewis, Luke Zettlemoyer, and Marjan Ghazvininejad. 2023. https://doi.org/10.18653/v1/2023.findings-acl.564 In-context examples selection for machine translation . In Findings of the Association for Computational Linguistics: ACL 2023, pages 8857--8873, Toronto, Canada. Association for Computational Linguistics

    doi:10.18653/v1/2023.findings-acl.564 2023

  2. [2]

    Duarte Alves, Nuno Guerreiro, Jo \ a o Alves, Jos \'e Pombal, Ricardo Rei, Jos \'e de Souza, Pierre Colombo, and Andre Martins. 2023. https://doi.org/10.18653/v1/2023.findings-emnlp.744 Steering large language models for machine translation with finetuning and in-context learning . In Findings of the Association for Computational Linguistics: EMNLP 2023, ...

    doi:10.18653/v1/2023.findings-emnlp.744 2023

  3. [3]

    Seth Aycock, David Stap, Di Wu, Christof Monz, and Khalil Sima'an. 2024. https://arxiv.org/abs/2409.19151 Can llms really learn to translate a low-resource language from one grammar book? Preprint, arXiv:2409.19151

    arXiv 2024

  4. [4]

    Rachel Bawden and Fran c ois Yvon. 2023. https://aclanthology.org/2023.eamt-1.16/ Investigating the translation performance of a large multilingual language model: the case of BLOOM . In Proceedings of the 24th Annual Conference of the European Association for Machine Translation, pages 157--170, Tampere, Finland. European Association for Machine Translation

    work page 2023

  5. [5]

    Luca Beurer-Kellner, Marc Fischer, and Martin Vechev. 2023. https://doi.org/10.1145/3591300 Prompting is programming: A query language for large language models . Proc. ACM Program. Lang., 7(PLDI)

    doi:10.1145/3591300 2023

  6. [6]

    Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jeffrey Wu, Clemens Winter, Christopher Hesse, Mark Chen, Eric Sigler, Mateusz Litwin...

    work page 2020

  7. [7]

    Samuel Cahyawijaya, Holy Lovenia, and Pascale Fung. 2024. https://doi.org/10.18653/v1/2024.naacl-long.24 LLM s are few-shot in-context low-resource language learners . In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), pages 405--433, Me...

    doi:10.18653/v1/2024.naacl-long.24 2024

  8. [8]

    Pranjal Chitale, Jay Gala, and Raj Dabre. 2024. https://doi.org/10.18653/v1/2024.findings-acl.440 An empirical study of in-context learning in LLM s for machine translation . In Findings of the Association for Computational Linguistics: ACL 2024, pages 7384--7406, Bangkok, Thailand. Association for Computational Linguistics

    doi:10.18653/v1/2024.findings-acl.440 2024

Show all 50 references
  1. [9]

    Sara Court and Micha Elsner. 2024. https://doi.org/10.18653/v1/2024.wmt-1.125 Shortcomings of LLM s for low-resource translation: Retrieval and understanding are both the problem . In Proceedings of the Ninth Conference on Machine Translation, pages 1332--1354, Miami, Florida,...

    2024 doi

  2. [10]

    DeepSeek-AI, Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, and et. al. 2025. https://arxiv.org/abs/2501.12948 Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning . Preprint,...

    2025 arXiv

  3. [11]

    Yue Deng, Wenxuan Zhang, Sinno Jialin Pan, and Lidong Bing. 2024. https://arxiv.org/abs/2310.06474 Multilingual jailbreak challenges in large language models . Preprint, arXiv:2310.06474

    2024 arXiv

  4. [12]

    Qingxiu Dong, Lei Li, Damai Dai, Ce Zheng, Jingyuan Ma, Rui Li, Heming Xia, Jingjing Xu, Zhiyong Wu, Baobao Chang, Xu Sun, Lei Li, and Zhifang Sui. 2024. https://doi.org/10.18653/v1/2024.emnlp-main.64 A survey on in-context learning . In Proceedings of the 2024 Conference on E...

    2024 doi

  5. [13]

    Micha Elsner and Jordan Needle. 2023. https://doi.org/10.18653/v1/2023.sigmorphon-1.2 Translating a low-resource language using GPT -3 and a human-readable dictionary . In Proceedings of the 20th SIGMORPHON workshop on Computational Research in Phonetics, Phonology, and Morpho...

    2023 doi

  6. [14]

    Marco Forni. 2013. Dizionario italiano -- ladino gardenese / Dizioner ladino gardenese -- italiano. Istitut Ladin Micurá de Rü, San Martin de Tor

    2013

  7. [15]

    Samuel Frontull and Georg Moser. 2024. https://doi.org/10.18653/v1/2024.loresmt-1.13 Rule-based, neural and LLM back-translation: Comparative insights from a variant of L adin . In Proceedings of the Seventh Workshop on Technologies for Machine Translation of Low-Resource Lang...

    2024 doi

  8. [16]

    Yuan Gao, Ruili Wang, and Feng Hou. 2024. https://doi.org/10.1145/3700410.3702123 How to design translation prompts for chatgpt: An empirical study . In Proceedings of the 6th ACM International Conference on Multimedia in Asia Workshops, MMAsia '24 Workshops, New York, NY, USA...

    2024

  9. [17]

    Jiatao Gu, Hany Hassan, Jacob Devlin, and Victor O.K. Li. 2018. https://doi.org/10.18653/v1/N18-1032 Universal neural machine translation for extremely low resource languages . In Proceedings of the 2018 Conference of the North A merican Chapter of the Association for Computat...

    2018 doi

  10. [18]

    Ping Guo, Yubing Ren, Yue Hu, Yunpeng Li, Jiarui Zhang, Xingsheng Zhang, and Heyan Huang. 2024. https://aclanthology.org/2024.lrec-main.1362/ Teaching large language models to translate on low-resource languages with textbook prompting . In Proceedings of the 2024 Joint Intern...

    2024

  11. [19]

    Amr Hendy, Mohamed Abdelrehim, Amr Sharaf, Vikas Raunak, Mohamed Gabr, Hitokazu Matsushita, Young Jin Kim, Mohamed Afify, and Hany Hassan Awadalla. 2023. https://arxiv.org/abs/2302.09210 How good are gpt models at machine translation? a comprehensive evaluation . Preprint, arX...

    2023 arXiv

  12. [20]

    Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, et al. 2024. Gpt-4o system card. arXiv preprint arXiv:2410.21276

    2024 arXiv

  13. [21]

    Aaron Jaech, Adam Kalai, Adam Lerer, Adam Richardson, Ahmed El-Kishky, Aiden Low, Alec Helyar, Aleksander Madry, Alex Beutel, Alex Carney, et al. 2024. Openai o1 system card. arXiv preprint arXiv:2412.16720

    2024 arXiv

  14. [22]

    u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt\

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich K\" u ttler, Mike Lewis, Wen-tau Yih, Tim Rockt\" a schel, Sebastian Riedel, and Douwe Kiela. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. In Proc...

    2020

  15. [23]

    Shushen Manakhimova, Eleftherios Avramidis, Vivien Macketanz, Ekaterina Lapshinova-Koltunski, Sergei Bagdasarov, and Sebastian M \"o ller. 2023. https://doi.org/10.18653/v1/2023.wmt-1.23 Linguistically motivated evaluation of the 2023 state-of-the-art machine translation: Can ...

    2023 doi

  16. [24]

    Rapha \"e l Merx, Aso Mahmudi, Katrina Langford, Leo Alberto de Araujo, and Ekaterina Vylomova. 2024. https://aclanthology.org/2024.eurali-1.1/ Low-resource machine translation through retrieval-augmented LLM prompting: A study on the M ambai language . In Proceedings of the 2...

    2024

  17. [25]

    Kelleher, and Andy Way

    Yasmin Moslem, Rejwanul Haque, John D. Kelleher, and Andy Way. 2023. https://aclanthology.org/2023.eamt-1.22/ Adaptive machine translation with large language models . In Proceedings of the 24th Annual Conference of the European Association for Machine Translation, pages 227--...

    2023

  18. [26]

    NLLB Team , Marta R. Costa-juss \`a , James Cross, Onur C elebi, Maha Elbayad, Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht, Jean Maillard, Anna Sun, Skyler Wang, Guillaume Wenzek, Al Youngblood, Bapi Akula, Loic Barrault, Gabriel Mejia Gonzalez,...

    2024

  19. [27]

    Vikt \'o ria Ondrejov \'a and Marek S uppa. 2024. https://doi.org/10.18653/v1/2024.vardial-1.11 Can LLM s handle low-resource dialects? a case study on translation and common sense reasoning in s ari s . In Proceedings of the Eleventh Workshop on NLP for Similar Languages, Var...

    2024 doi

  20. [28]

    Matt Post. 2018. https://www.aclweb.org/anthology/W18-6319 A call for clarity in reporting BLEU scores . In Proceedings of the Third Conference on Machine Translation: Research Papers, pages 186--191, Belgium, Brussels. Association for Computational Linguistics

    2018

  21. [29]

    Ratish Puduppully, Anoop Kunchukuttan, Raj Dabre, Ai Ti Aw, and Nancy Chen. 2023. https://doi.org/10.18653/v1/2023.emnlp-main.279 D eco MT : Decomposed prompting for machine translation between related languages using large language models . In Proceedings of the 2023 Conferen...

    2023 doi

  22. [30]

    Laria Reynolds and Kyle McDonell. 2021. https://doi.org/10.1145/3411763.3451760 Prompt programming for large language models: Beyond the few-shot paradigm . In Extended Abstracts of the 2021 CHI Conference on Human Factors in Computing Systems, CHI EA '21, New York, NY, USA. A...

    2021

  23. [31]

    Mortensen, and Graham Neubig

    Nathaniel Robinson, Perez Ogayo, David R. Mortensen, and Graham Neubig. 2023. https://doi.org/10.18653/v1/2023.wmt-1.40 C hat GPT MT : Competitive for high- (but not low-) resource languages . In Proceedings of the Eighth Conference on Machine Translation, pages 392--418, Sing...

    2023 doi

  24. [32]

    Ohad Rubin, Jonathan Herzig, and Jonathan Berant. 2022. https://doi.org/10.18653/v1/2022.naacl-main.191 Learning to retrieve prompts for in-context learning . In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: ...

    2022 doi

  25. [33]

    Barbara Scalvini, Iben Nyholm Debess, Annika Simonsen, and Hafsteinn Einarsson. 2025. https://aclanthology.org/2025.nodalida-1.62/ Rethinking low-resource MT: the surprising effectiveness of fine-tuned multilingual models in the LLM age . In Proceedings of the Joint 25th Nordi...

    2025

  26. [34]

    Rico Sennrich, Barry Haddow, and Alexandra Birch. 2016. https://doi.org/10.18653/v1/P16-1009 Improving neural machine translation models with monolingual data . In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), ...

    2016 doi

  27. [35]

    Peng Shu, Junhao Chen, Zhengliang Liu, Hui Wang, Zihao Wu, Tianyang Zhong, Yiwei Li, Huaqin Zhao, Hanqi Jiang, Yi Pan, Yifan Zhou, Constance Owl, Xiaoming Zhai, Ninghao Liu, Claudio Saunt, and Tianming Liu. 2024. https://arxiv.org/abs/2411.11295 Transcending language boundarie...

    2024 arXiv

  28. [36]

    David Stap, Eva Hasler, Bill Byrne, Christof Monz, and Ke Tran. 2024. https://doi.org/10.18653/v1/2024.acl-long.336 The fine-tuning paradox: Boosting translation quality without sacrificing LLM abilities . In Proceedings of the 62nd Annual Meeting of the Association for Comput...

    2024 doi

  29. [37]

    Cagri Toraman. 2024. https://doi.org/10.18653/v1/2024.mrl-1.3 Adapting open-source generative large language models for low-resource languages: A case study for T urkish . In Proceedings of the Fourth Workshop on Multilingual Representation Learning (MRL 2024), pages 30--44, M...

    2024 doi

  30. [38]

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, Aurelien Rodriguez, Armand Joulin, Edouard Grave, and Guillaume Lample. 2023. https://arxiv.org/abs/2302.13971 Llama:...

    2023 arXiv

  31. [39]

    Giovanni Valer, Nicolò Penzo, and Jacopo Staiano. 2024. Nesciun lengaz lascià endò: M achine translation for F assa L adin. In Proceedings of the 10th Italian Conference on Computational Linguistics, Pisa, Italy. CEUR-ws.org

    2024

  32. [40]

    Inacio Vieira, Will Allred, S \'e amus Lankford, Sheila Castilho, and Andy Way. 2024. https://aclanthology.org/2024.amta-research.20/ How much data is enough data? fine-tuning large language models for in-house translation: Performance evaluation across multiple dataset sizes ...

    2024

  33. [41]

    David Vilar, Markus Freitag, Colin Cherry, Jiaming Luo, Viresh Ratnakar, and George Foster. 2023. https://doi.org/10.18653/v1/2023.acl-long.859 Prompting P a LM for translation: Assessing strategies and performance . In Proceedings of the 61st Annual Meeting of the Association...

    2023 doi

  34. [42]

    Chi, Quoc V

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia, Ed H. Chi, Quoc V. Le, and Denny Zhou. 2022. Chain-of-thought prompting elicits reasoning in large language models. In Proceedings of the 36th International Conference on Neural Information Processi...

    2022

  35. [43]

    Sohee Yang, Elena Gribovskaya, Nora Kassner, Mor Geva, and Sebastian Riedel. 2024. https://doi.org/10.18653/v1/2024.acl-long.550 Do large language models latently perform multi-hop reasoning? In Proceedings of the 62nd Annual Meeting of the Association for Computational Lingui...

    2024 doi

  36. [44]

    Yixin Ye, Zhen Huang, Yang Xiao, Ethan Chern, Shijie Xia, and Pengfei Liu. 2025. https://arxiv.org/abs/2502.03387 Limo: Less is more for reasoning . Preprint, arXiv:2502.03387

    2025 arXiv

  37. [45]

    Zheng-Xin Yong, Cristina Menghini, and Stephen H. Bach. 2024. https://arxiv.org/abs/2310.02446 Low-resource languages jailbreak gpt-4 . Preprint, arXiv:2310.02446

    2024 arXiv

  38. [46]

    Zheng Xin Yong, Hailey Schoelkopf, Niklas Muennighoff, Alham Fikri Aji, David Ifeoluwa Adelani, Khalid Almubarak, M Saiful Bari, Lintang Sutawika, Jungo Kasai, Ahmed Baruwa, Genta Winata, Stella Biderman, Edward Raff, Dragomir Radev, and Vassilina Nikoulina. 2023. https://doi....

    2023 doi

  39. [47]

    Yi Zeng, Hongpeng Lin, Jingwen Zhang, Diyi Yang, Ruoxi Jia, and Weiyan Shi. 2024. https://doi.org/10.18653/v1/2024.acl-long.773 How johnny can persuade LLM s to jailbreak them: Rethinking persuasion to challenge AI safety by humanizing LLM s . In Proceedings of the 62nd Annual...

    2024 doi

  40. [48]

    Biao Zhang, Barry Haddow, and Alexandra Birch. 2023. Prompting large language model for machine translation: a case study. In Proceedings of the 40th International Conference on Machine Learning, ICML'23. JMLR.org

    2023

  41. [49]

    Chen Zhang, Xiao Liu, Jiuheng Lin, and Yansong Feng. 2024. https://doi.org/10.18653/v1/2024.findings-acl.519 Teaching large language models an unseen language on the fly . In Findings of the Association for Computational Linguistics: ACL 2024, pages 8783--8800, Bangkok, Thaila...

    2024 doi

  42. [50]

    Dawei Zhu, Pinzhen Chen, Miaoran Zhang, Barry Haddow, Xiaoyu Shen, and Dietrich Klakow. 2024. https://doi.org/10.18653/v1/2024.emnlp-main.24 Fine-tuning large language models to translate: Will a touch of noisy data in misaligned languages suffice? In Proceedings of the 2024 C...

    2024 doi

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.