@inproceedings{vamshi-yang-2026-protocode,
title = "Protocode: Prototype-Driven Interpretability for Code Generation in {LLM}s",
author = "Vamshi, Bodla Krishna and
Yang, Haizhao",
editor = "Zhang, Yue and
Xu, Ruifeng and
Purwarianti, Ayu and
Rong, Tong and
Suzuki, Jun",
booktitle = "Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the {A}sia-Pacific Chapter of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.aacl-long.20/",
pages = "386--407",
ISBN = "979-8-89176-439-2",
abstract = "Large Language Models (LLMs) are now widely used for tasks such as summarization, question answering, and speech-to-text translation, and increasingly for code generation, with tools such as Cursor and Windsurf analyzing massive code repositories and recommending changes. Big tech companies have acknowledged growing reliance on LLM-generated code. Although this improves developer productivity, it also raises the risk of suboptimal and insecure code. We introduce a retrieval framework for code generation that connects geometry-aware prototype sampling with syntax-grounded interpretability analysis. By conditioning AST-based explanations on retrieved demonstrations, our method evaluates retrieval quality through both functional correctness and interpretability alignment. In our experiments, high-quality in-context learning (ICL) demonstrations make outputs easier to interpret and improve P@10, whereas poorly chosen demonstrations reduce P@10 relative to the base model. Overall, our approach highlights the importance of efficient sampling strategies for ICL, which can affect model performance on any given task."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="vamshi-yang-2026-protocode">
<titleInfo>
<title>Protocode: Prototype-Driven Interpretability for Code Generation in LLMs</title>
</titleInfo>
<name type="personal">
<namePart type="given">Bodla</namePart>
<namePart type="given">Krishna</namePart>
<namePart type="family">Vamshi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haizhao</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayu</namePart>
<namePart type="family">Purwarianti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Rong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-439-2</identifier>
</relatedItem>
<abstract>Large Language Models (LLMs) are now widely used for tasks such as summarization, question answering, and speech-to-text translation, and increasingly for code generation, with tools such as Cursor and Windsurf analyzing massive code repositories and recommending changes. Big tech companies have acknowledged growing reliance on LLM-generated code. Although this improves developer productivity, it also raises the risk of suboptimal and insecure code. We introduce a retrieval framework for code generation that connects geometry-aware prototype sampling with syntax-grounded interpretability analysis. By conditioning AST-based explanations on retrieved demonstrations, our method evaluates retrieval quality through both functional correctness and interpretability alignment. In our experiments, high-quality in-context learning (ICL) demonstrations make outputs easier to interpret and improve P@10, whereas poorly chosen demonstrations reduce P@10 relative to the base model. Overall, our approach highlights the importance of efficient sampling strategies for ICL, which can affect model performance on any given task.</abstract>
<identifier type="citekey">vamshi-yang-2026-protocode</identifier>
<location>
<url>https://aclanthology.org/2026.aacl-long.20/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>386</start>
<end>407</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Protocode: Prototype-Driven Interpretability for Code Generation in LLMs
%A Vamshi, Bodla Krishna
%A Yang, Haizhao
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-439-2
%F vamshi-yang-2026-protocode
%X Large Language Models (LLMs) are now widely used for tasks such as summarization, question answering, and speech-to-text translation, and increasingly for code generation, with tools such as Cursor and Windsurf analyzing massive code repositories and recommending changes. Big tech companies have acknowledged growing reliance on LLM-generated code. Although this improves developer productivity, it also raises the risk of suboptimal and insecure code. We introduce a retrieval framework for code generation that connects geometry-aware prototype sampling with syntax-grounded interpretability analysis. By conditioning AST-based explanations on retrieved demonstrations, our method evaluates retrieval quality through both functional correctness and interpretability alignment. In our experiments, high-quality in-context learning (ICL) demonstrations make outputs easier to interpret and improve P@10, whereas poorly chosen demonstrations reduce P@10 relative to the base model. Overall, our approach highlights the importance of efficient sampling strategies for ICL, which can affect model performance on any given task.
%U https://aclanthology.org/2026.aacl-long.20/
%P 386-407
Markdown (Informal)
[Protocode: Prototype-Driven Interpretability for Code Generation in LLMs](https://aclanthology.org/2026.aacl-long.20/) (Vamshi & Yang, AACL-IJCNLP 2026)
ACL
- Bodla Krishna Vamshi and Haizhao Yang. 2026. Protocode: Prototype-Driven Interpretability for Code Generation in LLMs. In Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers), pages 386–407, Hengqin, China. Association for Computational Linguistics.