Teaching LLMs a Low-Resource Language: Enhancing Code Completion in Pharo

• URL: http://arxiv.org/abs/2607.04939v1
• Date: Mon, 06 Jul 2026 11:11:18 GMT
• Title: Teaching LLMs a Low-Resource Language: Enhancing Code Completion in Pharo
• Abstract summary: Large Language Models (LLMs) unlocked new possibilities in automated code writing, becoming the backbone of most code completion tools.<n>LLMs excel in mainstream languages, but they often lack support for the so-called low-resource languages where training data is scarce.<n>A concrete example is Pharo, a Smalltalk-inspired language whose IDE currently offers only single-token completion.
• Score: 51.12236828558682
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large Language Models (LLMs) unlocked new possibilities in automated code writing, becoming the backbone of most code completion tools. While LLMs excel in mainstream languages, they often lack support for the so-called low-resource languages where training data is scarce. As a result, these languages lag behind in the quality of code completion tooling available to their communities. A concrete example is Pharo, a Smalltalk-inspired language whose IDE currently offers only single-token completion. In this work, we report on our experience bringing LLM-based code completion to Pharo. First, we describe an end-to-end pipeline that combines Pharo-specific data curation, continued pre-training and fine-tuning of open code LLMs. Second, we introduce a set of Pharo code completion benchmarks designed to evaluate whether models (i) learn Pharo's syntax and (ii) accurately complete masked Pharo code from real-world GitHub repositories. Third, we show empirically that Pharo-specialized models substantially outperform their original base checkpoints and also exceed the accuracy of substantially larger code LLMs on Pharo completion. Overall, our case study demonstrates the feasibility of bringing strong LLM-based code completion to low-resource programming languages, with models small enough to provide ``real-time'' in-IDE support.

Related papers

• No Resource, No Benchmarks, No Problem? Evaluating and Improving LLMs for Code Generation in No-Resource Languages [50.536461876371526]
  Large Language Models (LLMs) have advanced the automation of software engineering tasks.<n>Most research in this area has focused on high-resource languages, such as Python or Java, which benefit from abundant training data.<n>No-resource languages for which LLMs have seen virtually no training data remain largely unstudied.
  arXiv  Detail & Related papers  (2026-06-15T15:08:55Z)
• LLM-JEPA: Large Language Models Meet Joint Embedding Predictive Architectures [50.494504099850325]
  Large Language Model (LLM) pretraining, finetuning, and evaluation rely on input-space reconstruction and generative capabilities.<n>Yet, it has been observed in vision that embedding-space training objectives, e.g., with Joint Embedding Predictive Architectures (JEPAs), are far superior to their input-space counterpart.
  arXiv  Detail & Related papers  (2025-09-11T03:03:57Z)
• The Fools are Certain; the Wise are Doubtful: Exploring LLM Confidence in Code Completion [4.215010577170175]
  We evaluate the confidence of Large Language Models (LLMs) when generating code by measuring code perplexity.<n>We find that strongly-typed languages exhibit lower perplexity than dynamically typed languages.<n> Perl appears universally high in perplexity, whereas Java appears low.
  arXiv  Detail & Related papers  (2025-08-22T06:51:13Z)
• I Can't Share Code, but I need Translation -- An Empirical Study on Code Translation through Federated LLM [3.9373541926236766]
  This study demonstrates that participants can collaboratively develop a FedLLM for efficient code translation.<n>Our findings indicate that FedLLM offers a collaborative approach to code translation and could serve as a promising direction for future research in this field.
  arXiv  Detail & Related papers  (2025-01-10T05:43:36Z)
• RepoQA: Evaluating Long Context Code Understanding [12.329233433333416]
  RepoQA is a benchmark to evaluate Large Language Models (LLMs) on long-context code understanding. RepoQA includes 500 code search tasks gathered from 50 popular repositories across 5 modern programming languages.
  arXiv  Detail & Related papers  (2024-06-10T05:15:30Z)
• MAP-Neo: Highly Capable and Transparent Bilingual Large Language Model Series [86.31735321970481]
  We open-source MAP-Neo, a bilingual language model with 7B parameters trained from scratch on 4.5T high-quality tokens. Our MAP-Neo is the first fully open-sourced bilingual LLM with comparable performance compared to existing state-of-the-art LLMs.
  arXiv  Detail & Related papers  (2024-05-29T17:57:16Z)
• If LLM Is the Wizard, Then Code Is the Wand: A Survey on How Code Empowers Large Language Models to Serve as Intelligent Agents [81.60906807941188]
  Large language models (LLMs) are trained on a combination of natural language and formal language (code) Code translates high-level goals into executable steps, featuring standard syntax, logical consistency, abstraction, and modularity.
  arXiv  Detail & Related papers  (2024-01-01T16:51:20Z)
• Knowledge Transfer from High-Resource to Low-Resource Programming Languages for Code LLMs [2.9242435458494445]
  This paper presents an effective approach for boosting the performance of Code LLMs on low-resource languages using semi-synthetic data. We apply this approach to generate tens of thousands of validated training items for Julia, Lua, OCaml, R, and Racket.
  arXiv  Detail & Related papers  (2023-08-19T03:19:01Z)
• Okapi: Instruction-tuned Large Language Models in Multiple Languages with Reinforcement Learning from Human Feedback [61.83548032416181]
  We present Okapi, the first system with instruction-tuned LLMs based on RLHF for multiple languages. Okapi introduces instruction and response-ranked data in 26 diverse languages to facilitate the experiments and development of future multilingual LLM research.
  arXiv  Detail & Related papers  (2023-07-29T18:01:46Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.