@inproceedings{chen-etal-2026-tipcoder,
title = "{T}ip{C}oder: Reinforcement Learning Boosted Test-time Instruction Proposer for Code Generation",
author = "Chen, Minyu and
Wu, Sihao and
Wu, Ling-I and
Qin, Song and
Li, Jingyang and
Ning, Lei and
Xue, Jianxin and
Li, Guoqiang",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.495/",
pages = "10363--10377",
ISBN = "979-8-89176-462-0",
abstract = "Test-time scaling for code generation typically explores the solution space by sampling multiple programs from a fixed instruction. We study a complementary direction: instance-level instruction-space exploration. Our observation is that many coding failures stem from missing constraints, overlooked edge cases, or misleading reasoning paths induced by the original prompt. To address this, we propose \textsc{TipCoder}, a test-time instruction proposer that generates problem-specific auxiliary tips before code synthesis. \textsc{TipCoder} distills multi-turn debugging trajectories into proactive guidance and further optimizes the Proposer with reinforcement learning using a marginal-utility reward. At inference time, it generates both a base solution and a tip-guided solution, and applies a Reward Model for post-hoc selection. This exploration-selection design allows tips to expose additional candidate potential while reducing regressions from unnecessary guidance. Across multiple code-generation benchmarks and target Code LLMs, \textsc{TipCoder} provides a robust instruction-level test-time scaling strategy, comparing favorably with stochastic sampling and generic prompt optimization baselines under a shared reward-model-based selection protocol."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chen-etal-2026-tipcoder">
<titleInfo>
<title>TipCoder: Reinforcement Learning Boosted Test-time Instruction Proposer for Code Generation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Minyu</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sihao</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ling-I</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Song</namePart>
<namePart type="family">Qin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jingyang</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lei</namePart>
<namePart type="family">Ning</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jianxin</namePart>
<namePart type="family">Xue</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guoqiang</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Test-time scaling for code generation typically explores the solution space by sampling multiple programs from a fixed instruction. We study a complementary direction: instance-level instruction-space exploration. Our observation is that many coding failures stem from missing constraints, overlooked edge cases, or misleading reasoning paths induced by the original prompt. To address this, we propose TipCoder, a test-time instruction proposer that generates problem-specific auxiliary tips before code synthesis. TipCoder distills multi-turn debugging trajectories into proactive guidance and further optimizes the Proposer with reinforcement learning using a marginal-utility reward. At inference time, it generates both a base solution and a tip-guided solution, and applies a Reward Model for post-hoc selection. This exploration-selection design allows tips to expose additional candidate potential while reducing regressions from unnecessary guidance. Across multiple code-generation benchmarks and target Code LLMs, TipCoder provides a robust instruction-level test-time scaling strategy, comparing favorably with stochastic sampling and generic prompt optimization baselines under a shared reward-model-based selection protocol.</abstract>
<identifier type="citekey">chen-etal-2026-tipcoder</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.495/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>10363</start>
<end>10377</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T TipCoder: Reinforcement Learning Boosted Test-time Instruction Proposer for Code Generation
%A Chen, Minyu
%A Wu, Sihao
%A Wu, Ling-I
%A Qin, Song
%A Li, Jingyang
%A Ning, Lei
%A Xue, Jianxin
%A Li, Guoqiang
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F chen-etal-2026-tipcoder
%X Test-time scaling for code generation typically explores the solution space by sampling multiple programs from a fixed instruction. We study a complementary direction: instance-level instruction-space exploration. Our observation is that many coding failures stem from missing constraints, overlooked edge cases, or misleading reasoning paths induced by the original prompt. To address this, we propose TipCoder, a test-time instruction proposer that generates problem-specific auxiliary tips before code synthesis. TipCoder distills multi-turn debugging trajectories into proactive guidance and further optimizes the Proposer with reinforcement learning using a marginal-utility reward. At inference time, it generates both a base solution and a tip-guided solution, and applies a Reward Model for post-hoc selection. This exploration-selection design allows tips to expose additional candidate potential while reducing regressions from unnecessary guidance. Across multiple code-generation benchmarks and target Code LLMs, TipCoder provides a robust instruction-level test-time scaling strategy, comparing favorably with stochastic sampling and generic prompt optimization baselines under a shared reward-model-based selection protocol.
%U https://aclanthology.org/2026.findings-emnlp.495/
%P 10363-10377
Markdown (Informal)
[TipCoder: Reinforcement Learning Boosted Test-time Instruction Proposer for Code Generation](https://aclanthology.org/2026.findings-emnlp.495/) (Chen et al., Findings 2026)
ACL
- Minyu Chen, Sihao Wu, Ling-I Wu, Song Qin, Jingyang Li, Lei Ning, Jianxin Xue, and Guoqiang Li. 2026. TipCoder: Reinforcement Learning Boosted Test-time Instruction Proposer for Code Generation. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 10363–10377, Budapest, Hungary. Association for Computational Linguistics.