Large Language Models and Language Server Protocol: a match made in context
Source
arxiv.org
Author
Alessandro Schena, Ilgiz Mustafin, Julia Kotovich
Date
Context
Eiffel-tools puts a formal checker into the loop when a LLMA large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.Full definition → generates or repairs Eiffel programs. The tool supplies language and project context windowThe maximum amount of text a model can consider at once — its working memory for the current conversation or task.Full definition →, examines the generated output, and manages further attempts when verification fails. Its verifier checks annotated programs against formal specifications, giving the revision process a defined criterion for acceptance. Across two public datasets and three models, the authors report fixing 76% to 95% of bugs, depending on the model and prompts. Those results concern the evaluated Eiffel tasks; the useful mechanism is coupling generated repairs to verification feedback, with success also depending on how many attempts are allowed. Paper.
Terms in this piece · Glossary
LLM — A large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
articleAutomated Vulnerability Injection in Smart Contracts Using Large Language ModelsLuca Migliaccio, Roberto Natella, Naghmeh Ivaki, Nuno Laranjeiro, Marco Vieira
articleRFCLLM: Evaluating LLMs' Reasoning Ability of Network Protocol State MachinesAnqi Chen, Dan Goldwasser, Cristina Nita-Rotaru
articleDo LLMs Make More Mistakes If They Do Not Believe the Input Data?Peter Kochelka, Ale\v{s} Manuel Pap\'a\v{c}ek, Vojt\v{e}ch Dvo\v{r}\'ak, Ond\v{r}ej Du\v{s}ek