Pith. sign in

REVIEW 3 major objections 6 minor 33 references

Can LLMs Help Create Grammar?: Automating Grammar Creation for Endangered Languages with In-Context Learning

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read An LLM with in-context learning can generate coherent grammar for endangered Moklen from just a dictionary and parallel sentences, with 86 of 100 lexical entries judged accurate.

desk verdict A genuinely new task with a clean pipeline, but the evaluation never checks whether the generated grammar actually parses, so the central claim is not yet supported. read the letter →

arxiv 2412.10960 v1 pith:V37T566Q submitted 2024-12-14 cs.CL

Piyapath T Spencer , Nanthipat Kongborrirak This is my paper · ORCID
classification cs.CL
keywords endangeredlanguagesMoklenin-contextlearninggrammarinductionLexical-FunctionalXLElow-resourceNLPlanguagedocumentation
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

This paper tries to show that a large language model, prompted with only a Moklen-to-English dictionary and a handful of parallel sentences, can write a formal grammar for Moklen in the XLE formalism, plus lexical entries for words it never saw in sentences. The authors use Moklen, an endangered Austronesian language of southern Thailand with fewer than 1,000 speakers, as a test case because it is isolating and lacks inflectional morphology. They report that the best prompt combination, which adds tokenised dictionary senses and the model's own self-explanation, raises Moklen-to-English BERTScore F1 to 0.7110 when the generated grammar is included, and that 86 of 100 generated lexical entries are accurate against dictionary definitions. If the claim holds, grammar writing for endangered languages could start from the kinds of data linguists already collect, without pretraining a model or hiring a grammar engineer. The paper itself notes that English grammatical bias and the isolating typology of Moklen may have made the task easier.

What carries the argument

The mechanism is an in-context learning pipeline with five stages: a dictionary-based tokeniser using longest-match segmentation; sense mapping that attaches English glosses to each Moklen token; concatenation of word-by-word glosses into a string that mirrors Moklen word order; prompting the LLM with these materials plus XLE documentation and optional self-explanation; and finally using the generated grammar rules to prompt lexical entries for unseen dictionary words. The load-bearing formal object is the XLE grammar, a computational implementation of Lexical-Functional Grammar whose rules and lexical entries pair constituent structure with functional structure. The pipeline's effect is to give the LLM enough aligned local evidence to infer word order, parts of speech, and lexical schemata by analogy with English, without any Moklen text in pretraining.

What would settle it

Have a second linguist with no access to Spencer (2022) independently write a reference grammar of Moklen from primary fieldwork, then compare it rule-by-rule with the gold standard and with the generated XLE grammar; if the independent grammar disagrees on core word order, parts of speech, or lexical senses, the evaluation baseline is unreliable. A more computational check is to feed the generated grammar to the XLE parser on the 40 held-out sentences and require a successful parse and f-structure for sentences the gold grammar parses.

Watch

Extended reading notes

Core claim

The central discovery is that GPT-4o-mini, a small API-based LLM, can induce a coherent XLE grammar and useful lexical entries for a language it has not been trained on, provided its prompt contains a bilingual dictionary, tokenised and sense-mapped parallel sentences, and XLE documentation. The authors find that the most effective context for translation is tokenised dictionary senses plus concatenated word-by-word meanings plus self-explanation (TD+C+S), and that adding grammar to the prompt improves translation scores across most settings. In direct evaluation, 86 out of 100 lexical entries generated for Moklen words absent from the bitext were judged accurate and coherent, while the main observed error type was missing secondary senses of polysemous words and an over-application of English parts of speech such as determiners. This is framed as evidence that LLMs can assist language documentation, not replace it.

Load-bearing premise

The accuracy judgements presuppose that the unpublished reference grammar of Moklen used as the gold standard is correct and complete, and that the dictionary covers every word appearing in the bitext; if either premise fails, the reported accuracy numbers stop supporting the conclusions.

Editorial extensions

If this is right

  • If correct, grammar construction for endangered languages could begin from a bilingual wordlist and a few dozen recorded sentences, greatly lowering the cost of formal grammar writing.
  • Generated lexical entries can extend a dictionary to words not present in any sentence corpus, since 86 of 100 entries in the study were judged accurate.
  • Adding the generated grammar to the prompt improves translation quality, suggesting grammar rules and lexical resources can feed back into machine translation.
  • The same prompting strategy may transfer to other formal grammar frameworks beyond XLE, such as dependency grammars or HPSG, as the paper suggests.

Reading between the lines

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

  • A testable extension is to run the identical pipeline on a morphologically rich endangered language; the paper's limitation note suggests the isolating typology of Moklen may be doing much of the work, so success on agreement-heavy languages would be a stronger test.
  • The BERTScore gain after adding grammar could partly reflect the evaluator's sensitivity to content-word overlap rather than true syntactic understanding, so a parse-based evaluation of the generated XLE grammar against independently collected sentences would separate the two.
  • Hallucinated categories that the model inserted into Moklen grammar, such as determiners, might function as hypotheses for field linguists to check, turning model errors into a discovery aid.
  • Dictionary completeness is the quiet precondition: the tokeniser fails if the bitext contains words absent from the dictionary, so field applications would need a morphological guesser or a mechanism to flag out-of-vocabulary items.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes using GPT-4o-mini with in-context learning to generate formal XLE (LFG) grammar rules and lexical entries for Moklen, an endangered Austronesian language, using only a bilingual dictionary and a small parallel bitext. The method consists of dictionary-based tokenisation, sense mapping, concatenation of per-word translations, and a prompt that varies several context types (bitext, tokenisation, dictionary, concatenated sentence, XLE example, self-explanation). Translation BERTScore is used to select among 48 context combinations; grammar rules are evaluated qualitatively against a gold standard constructed from Spencer (2022); lexical entries for 100 words not in the bitext are manually assessed against dictionary definitions. The paper reports that the TD+C+S context combination achieves BERTScore F1 of 0.7110 for Moklen-to-English translation when an XLE gold-standard grammar is added to the prompt, and that 86 of 100 generated lexical entries are accurate.

Significance. If properly validated, the approach would be a valuable contribution to language documentation, where manual grammar engineering is a major bottleneck. The paper deserves credit for working with a real endangered language, deliberately using a small model, and for stating that data, code, and model generations are in the supplementary material. However, the current evidence does not establish the central claim that the model generates a coherent XLE grammar: no parser is run on the generated grammar, the grammar gold standard is an unpublished manuscript by the first author, and the headline translation number appears to be selected on the test set. These are not merely presentation issues; they affect the validity of the main results. The idea is promising, but the paper needs substantially stronger validation before publication.

major comments (3)
  1. [Section 6.1] The reported peak BERTScore F1 of 0.7110 for TD+C+S is obtained after exploring 48 context combinations, and the evaluation appears to use the same 40-sentence parallel set described in Section 5.2 as the basis for selecting the best context. If so, the headline number is an in-sample selection optimum, and the comparison among contexts does not support the claim that TD+C+S is the best setting. Please clarify the exact split between the ablation set and the evaluation set, and ideally re-run context selection on a development set with a held-out test set.
  2. [Sections 5.2 and 6.2] The grammar-rule evaluation relies on a gold standard that the authors 'attempt to create one based on Spencer (2022)', an unpublished manuscript by the first author. That source is neither available to readers nor independently verified, so the accuracy judgements for grammar rules are non-reproducible and potentially circular. More importantly, the paper never runs the XLE parser on the generated grammar and lexical entries, even though 'coherent XLE grammar' is the central claim. The translation improvements in Section 6.1 are obtained by adding the 'XLE gold standard grammar' to the prompt, not the grammar produced by the model, so those results do not validate the generated grammar. Section 6.2 also reports no quantitative accuracy or coverage score for the grammar rules.
  3. [Section 6.3] The 86/100 lexical-entry accuracy is assessed by comparing generated entries to dictionary definitions that were also provided as context in the prompt (Sections 4.3 and 5.2). This makes the evaluation partly circular: high agreement with the input dictionary does not demonstrate that the model has learned Moklen grammar or generalised beyond the provided lexicon. In addition, the manual assessment is reported without a blind protocol or inter-annotator agreement, so the 86% figure is difficult to interpret.
minor comments (6)
  1. [Abstract] The abstract contains typographical and grammatical errors, including 'Yes!' at the start and 'We takes Moklen as a case study'; please copyedit the entire manuscript.
  2. [Sections 5.1 and 6.1] There are unresolved cross-references 'See ?? for full details' in Sections 5.1 and 6.1, and several figure captions are malformed or incomplete (for example, Figure 3 lacks a clear legend and readable axis labels).
  3. [Section 5.2] The paper lists BLEU, ROUGE, METEOR, chrF, and BERTScore as evaluation metrics, but Section 6.1 reports only BERTScore; please report the other metrics or explicitly justify their omission.
  4. [Section 4.1] The tokeniser depends on the assumption that every word in the Moklen bitext appears in the dictionary; please verify this against the actual data and report the number of out-of-dictionary tokens, if any.
  5. [Section 6.2] The section title is 'Grammar Rules: Accuracy and Completeness', but no accuracy numbers are given; either add a quantitative evaluation or rename the section to reflect the qualitative discussion.
  6. [Limitation] The limitation section addresses typological generality, but it does not mention the lack of parser validation or the reliance on an unpublished gold standard; both should be acknowledged.

Circularity Check

2 steps flagged · score 4.0 of 10

Lexical-entry accuracy is assessed against the dictionary definitions already given in the prompt, and grammar-rule accuracy is graded against a gold standard built from the first author's unpublished manuscript; the central XLE-parsing claim is unevaluated but the derivation is not fully circular.

  1. fitted input called prediction [Section 5.2 (Lexical Entry and Schemata Accuracy) and Section 6.3 (Lexical Entries: Accuracy and Coherence)]
    "The model will generate lexical entries for 100 words from the Moklen dictionary that do not appear in the bitext, primarily based on the grammar rules produced by the model and the information provided in the dictionary. ... We generated lexical entries for 100 Moklen words not present in the bitext and assessed their accuracy by comparing them to existing dictionary definitions. Out of these, 86 entries were deemed accurate and coherent."

    The dictionary definitions are the input used to generate the lexical entries, and the same dictionary definitions are the reference against which accuracy is measured. The 86/100 accuracy therefore largely measures whether the model reformatted information that was already in the prompt; it is not an independent prediction of new lexical knowledge. The score is partly forced by construction rather than by the generated grammar's quality.

  2. self citation load bearing [Section 5.2 (Grammar Rule Accuracy)]
    "Never before had Moklen been developed its grammar using XLE. Hence, there is no XLE grammar for Moklen; we attempt to create one based on Spencer (2022) and served as a gold standard."

    The gold standard used to judge whether the LLM's grammar rules are 'coherent' is an unpublished manuscript by the paper's first author. The evaluation of the central grammar-generation claim therefore rests on a self-authored, unverified standard: the judgment reduces to whether the LLM's output matches the authors' prior analysis, with no independent parser check or externally validated grammar to break the loop. This is not a formal tautology because the LLM never sees Spencer (2022), but the load-bearing evidence for the central claim is a self-citation.

full rationale

The paper's central claim is that gpt-4o-mini produces coherent XLE grammar and lexical entries for Moklen. The derivation chain is not fully circular: the generation step does use only the dictionary, parallel sentences, and XLE documentation, and the translation experiments are an independent (though weak) signal. However, the two direct evaluations of the generated outputs are both anchored to inputs or to the authors' own prior analysis. Lexical-entry accuracy compares the outputs to the dictionary definitions that were fed into the prompt, making the 86/100 figure a self-consistency check rather than an independent validation. Grammar-rule accuracy uses a gold standard 'attempt[ed]' from Spencer (2022), an unpublished manuscript by the first author, so the coherence verdict is not independently established. The headline BERTScore result (0.7110) is also reported for a condition that incorporates the XLE gold standard grammar rather than the generated grammar, so it does not directly verify the generated grammar. No experiment feeds the generated grammar into the XLE parser; that is a correctness and completeness risk, not itself circularity. Weighing these, the central ability claim has some independent content (translation improvements, manual coherence observations), but the most precise accuracy numbers reduce to matching prompt inputs or to a self-authored gold standard. Score 4 reflects partial circularity in the evaluation of the key outputs rather than a fully forced derivation.

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

The pipeline depends on unverified domain assumptions: the correctness and completeness of the Moklen dictionary and bitext, the isolating-language typology that makes word-level segmentation work, the validity of the author-built gold standard grammar, and the appropriateness of BERTScore as a quality metric. No numerical parameters are fitted, but the winning prompt context is selected from 48 candidates using the same test set that produces the headline score, which acts as a model-selection knob.

assumptions (5)
  • domain assumption Moklen is an isolating language with no inflectional morphology, so words can be segmented purely with a dictionary and longest-match tokenisation.
    Section 3 states Moklen lacks inflectional morphology and Section 4.1 relies on this to justify dictionary-based tokenisation.
  • domain assumption Every word in the bitext appears in the Moklen-English dictionary, so the tokeniser can segment all sentences.
    Section 4.1: 'all words in the Moklen sentences from the bitext exist in the dictionary.' Errors in dictionary coverage would propagate to grammar induction.
  • domain assumption The gold standard XLE grammar for Moklen, based on Spencer (2022), is a correct reference for evaluating generated grammar.
    Section 5.2: 'we attempt to create one based on Spencer (2022) and served as a gold standard.' This standard is an unpublished manuscript by one of the authors.
  • domain assumption BERTScore is a meaningful measure of translation quality for Moklen, and score differences reflect grammar quality.
    Section 7 argues BERTScore is preferred over word-form metrics for isolating languages; Section 6.1 uses BERTScore F1 as the headline metric. The paper itself notes BERTScore 'falls short in fully capturing the subtleties'.
  • domain assumption GPT-4o-mini has effectively zero prior knowledge of Moklen, so any correct output must come from in-context learning.
    Section 5.1: the no-context condition is used to 'validate that they have effectively learned zero Moklen during pretraining.' No direct evidence of this is given beyond poor zero-shot scores.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Can LLMs Help Create Grammar?: Automating Grammar Creation for Endangered Languages with In-Context Learning." pith.science (2026). https://pith.science/paper/V37T566Q

@misc{pith2026241210960,
  author       = {Pith},
  title        = {Pith review of: Can LLMs Help Create Grammar?: Automating Grammar Creation for Endangered Languages with In-Context Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/V37T566Q}},
  note         = {Machine review of arXiv:2412.10960}
}
read the original abstract

Yes! In the present-day documenting and preserving endangered languages, the application of Large Language Models (LLMs) presents a promising approach. This paper explores how LLMs, particularly through in-context learning, can assist in generating grammatical information for low-resource languages with limited amount of data. We takes Moklen as a case study to evaluate the efficacy of LLMs in producing coherent grammatical rules and lexical entries using only bilingual dictionaries and parallel sentences of the unknown language without building the model from scratch. Our methodology involves organising the existing linguistic data and prompting to efficiently enable to generate formal XLE grammar. Our results demonstrate that LLMs can successfully capture key grammatical structures and lexical information, although challenges such as the potential for English grammatical biases remain. This study highlights the potential of LLMs to enhance language documentation efforts, providing a cost-effective solution for generating linguistic data and contributing to the preservation of endangered languages.

Figures

Figures reproduced from arXiv: 2412.10960 by the authors.

Figure 1
Figure 1. The ϕ mapping (Dalrymple, 2023) from dif￾ferent c(onstituent)-structures of Moklen and English to the same f(unctional)-structure According to Butt et al. (1999), "a common set of linguistic principles and a commonly agreed upon set of grammatical analyses and features" unify linguistic insights across languages whilst ac￾knowledge the uniqueness of individual languages. The XLE (Xerox Linguistic Environment), de￾ve… view at source ↗
Figure 2
Figure 2. The approach requires each pair of sentences from bitext going through the process of tokenisation, sense [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. When instructing with grammar, the model [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: The improvement in scores across contexts [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: The clear improvement after intergrating dic [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 19 canonical work pages

  1. [1]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRING...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Akari Asai, Zexuan Zhong, Danqi Chen, Pang Wei Koh, Luke Zettlemoyer, Hannaneh Hajishirzi, and Wen tau Yih. 2024. https://arxiv.org/abs/2403.03187 Reliable, adaptable, and attributable language models with retrieval . Preprint, arXiv:2403.03187

    arXiv 2024

  4. [4]

    Satanjeev Banerjee and Alon Lavie. 2005. https://aclanthology.org/W05-0909 METEOR : An automatic metric for MT evaluation with improved correlation with human judgments . In Proceedings of the ACL Workshop on Intrinsic and Extrinsic Evaluation Measures for Machine Translation and/or Summarization , pages 65--72, Ann Arbor, Michigan. Association for Comput...

    2005

  5. [5]

    Gašper Beguš, Maksymilian Dąbkowski, and Ryan Rhodes. 2023. https://arxiv.org/abs/2305.00948 Large linguistic models: Analyzing theoretical linguistic abilities of llms . Preprint, arXiv:2305.00948

    arXiv 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...

    arXiv 2020

  7. [7]

    Miriam Butt, Helge Dyvik, Tracy Holloway King, Hiroshi Masuichi, and Christian Rohrer. 2002. https://aclanthology.org/W02-1503 The parallel grammar project . In COLING -02: Grammar Engineering and Evaluation

    work page 2002

  8. [8]

    Miriam Butt, Tracy Holloway King, Maria-Eugenia Niño, and Frederique Segond. 1999. A Grammar Writer's Cookbook. Stanford University Press

    work page 1999

Show all 33 references
  1. [9]

    Noam Chomsky and Robert Berwick. 2016. Why Only Us. MIT Press

    2016

  2. [10]

    Ann Copestake. 2002. Definitions of typed feature structures. In Stephan Oepen, Dan Flickinger, Jun-ichi Tsujii, and Hans Uszkoreit, editors, Collaborative Language Engineering, pages 227--230. CSLI Publications, Stanford, CA

    2002

  3. [11]

    Kaplan, Tracy Holloway King, John T

    Richard Crouch, Mary Dalrymple, Ronald M. Kaplan, Tracy Holloway King, John T. III Maxwell, and Paula S. Newman. 2011. https://ling.sprachwiss.uni-konstanz.de/pages/xle/doc/xle_toc.html. XLE Documentation

    2011

  4. [12]

    Mary Dalrymple, editor. 2023. https://doi.org/10.5281/zenodo.10037797 Handbook of Lexical Functional Grammar . Number 13 in Empirically Oriented Theoretical Morphology and Syntax. Language Science Press, Berlin

    2023 doi

  5. [13]

    Ethnologue. 2024. How many languages are endangered?

    2024

  6. [14]

    Kavita Ganesan. 2018. https://arxiv.org/abs/1803.01937 Rouge 2.0: Updated and improved measures for evaluation of summarization tasks . Preprint, arXiv:1803.01937

    2018 arXiv

  7. [15]

    Maxwell, III and Ronald M

    John T. Maxwell, III and Ronald M. Kaplan. The interface between phrasal and functional constraints. Computational Linguistics

  8. [16]

    Diego Miranda-Saavedra. 2024. Generative ai models and the quest for human-level artificial intelligence. Real World Data Science

    2024

  9. [17]

    Stephan Oepen, Kristina Toutanova, Stuart Shieber, Christopher Manning, Dan Flickinger, and Thorsten Brants. 2002. https://aclanthology.org/C02-2025 The L in GO redwoods treebank: Motivation and preliminary applications . In COLING 2002: The 17th International Conference on Co...

    2002

  10. [18]

    OpenAI. 2024. https://arxiv.org/abs/2303.08774 Gpt-4 technical report . Preprint, arXiv:2303.08774

    2024 arXiv

  11. [19]

    Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. https://doi.org/10.3115/1073083.1073135 B leu: a method for automatic evaluation of machine translation . In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, pages 311--3...

    2002

  12. [20]

    Forthcoming

    Pittayawat Pittayaporn and Songphan Choemprayong. Forthcoming. A proposal for a thai-based moklen orthography. Language Documentation and Conservation

  13. [21]

    Pittayawat Pittayaporn, Warunsiri Pornpottanamas, Daniel Loss, Songphan Choemprayong, Chengnan Zhang, Piyangkoon Thaweepol, Pongporn Prayongsap, Pornsuda Nawarak, Peerasak Sirisawasdi, and Leena Maluleem. 2022. Moklen-Thai-English Dictionary: A Pilot Version

    2022

  14. [22]

    Maja Popovi \'c . 2015. https://doi.org/10.18653/v1/W15-3049 chr F : character n-gram F -score for automatic MT evaluation . In Proceedings of the Tenth Workshop on Statistical Machine Translation, pages 392--395, Lisbon, Portugal. Association for Computational Linguistics

    2015 doi

  15. [23]

    Pritchett

    Bradley L. Pritchett. 1992. Grammatical Competence and Parsing Performance. The University of Chicago Press

    1992

  16. [24]

    Jane Simpson. 1991. Warlpiri Morpho-Syntax: A Lexicalist Approach. Kluwer Academic Publisher

    1991

  17. [25]

    Piyapath Spencer. 2024. https://aclanthology.org/2024.fieldmatters-1.4 Documenting endangered languages with L ang D oc: A wordlist-based system and a case study on M oklen . In Proceedings of the 3rd Workshop on NLP Applications to Field Linguistics (Field Matters 2024), page...

    2024

  18. [26]

    Piyapath T Spencer. 2022. Grammatical sketch of moklen verb phrase. Unpublished Manuscipt

    2022

  19. [27]

    Garrett Tanzer, Mirac Suzgun, Eline Visser, Dan Jurafsky, and Luke Melas-Kyriazi. 2024. https://openreview.net/forum?id=tbVWug9f2h A benchmark for learning to translate a new language from one grammar book

    2024

  20. [28]

    Gemini Team. 2024. https://arxiv.org/abs/2403.05530 Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context . Preprint, arXiv:2403.05530

    2024 arXiv

  21. [29]

    W3Techs. 2024. Usage statistics of content languages for websites

    2024

  22. [30]

    Zhao, Kelvin Guu, Adams Wei Yu, Brian Lester, Nan Du, Andrew M

    Jason Wei, Maarten Bosma, Vincent Y. Zhao, Kelvin Guu, Adams Wei Yu, Brian Lester, Nan Du, Andrew M. Dai, and Quoc V. Le. 2022. https://arxiv.org/abs/2109.01652 Finetuned language models are zero-shot learners . Preprint, arXiv:2109.01652

    2022 arXiv

  23. [31]

    Trimble, and Emily M

    Olga Zamaraeva, Chris Curtis, Guy Emerson, Antske Fokkens, Michael Goodman, Kristen Howell, T.J. Trimble, and Emily M. Bender. 2022. https://doi.org/10.15398/jlm.v10i1.292 20 years of the grammar matrix: cross-linguistic hypothesis testing of increasingly complex interactions ...

    2022 doi

  24. [32]

    Kexun Zhang, Yee Choi, Zhenqiao Song, Taiqi He, William Yang Wang, and Lei Li. 2024. https://aclanthology.org/2024.findings-acl.925 Hire a linguist!: Learning endangered languages in LLM s with in-context linguistic descriptions . In Findings of the Association for Computation...

    2024

  25. [33]

    Weinberger, and Yoav Artzi

    Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q. Weinberger, and Yoav Artzi. 2020. https://arxiv.org/abs/1904.09675 Bertscore: Evaluating text generation with bert . Preprint, arXiv:1904.09675

    2020 arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.