NES: An Instruction-Free, Low-Latency Next Edit Suggestion Framework Powered by Learned Historical Editing Trajectories
Not on the map. Why?
Abstract (arXiv)
Code editing is a frequent yet cognitively demanding task in software development. Existing AI-powered tools often disrupt developer flow by requiring explicit natural language instructions and suffer from high latency, limiting real-world usability. We present NES (Next Edit Suggestion), an instruction-free, low-latency code editing framework that leverages learned historical editing trajectories to implicitly capture developers' goals and coding habits. NES features a dual-model architecture: one model predicts the next edit location and the other generates the precise code change, both without any user instruction. Trained on our open-sourced SFT and DAPO datasets, NES achieves state-of-the-art performance (75.6% location accuracy, 27.7% exact match rate) while delivering suggestions in under 250ms. Deployed at Ant Group, NES serves over 20,000 developers through a seamless Tab-key interaction, achieving effective acceptance rates of 51.55% for location predictions and 43.44% for edits, demonstrating its practical impact in real-world development workflows.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 28 of its 59 references that are in the map. First posted on arXiv 4 Aug 2025.
Builds on
28 of 59 references.
References · 28
- 2022Training language models to follow instructions with human feedbackcited by 15,971
- 2021Evaluating Large Language Models Trained on Codecited by 8,320
- 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,810
- 2023Code Llama: Open Foundation Models for Codecited by 2,139
- 2022Fast Inference from Transformers via Speculative Decodingcited by 1,653
- 2020CodeBERT: A Pre-Trained Model for Programming and Natural Languagescited by 1,310
- 2025Kimi k1.5: Scaling Reinforcement Learning with LLMscited by 1,210
- 2023SGLang: Efficient Execution of Structured Language Model Programscited by 1,141
- 2023Accelerating Large Language Model Decoding with Speculative Samplingcited by 1,006
- 2021CodeT5: Identifier-aware Unified Pre-trained Encoder-Decoder Models for Code Understanding and Generationcited by 906
- 2022ReACC: A Retrieval-Augmented Code Completion Frameworkcited by 119
- 2024SWE-Bench+: Enhanced Coding Benchmark for LLMscited by 99
16 more
- 2025ACECODER: Acing Coder RL via Automated Test-Case Synthesiscited by 93
- 2025ACEBench: Who Wins the Match Point in Tool Usage?cited by 71
- 2020Multi-task Learning based Pre-trained Language Model for Code Completioncited by 58
- 2022CoditT5: Pretraining for Source Code and Natural Language Editingcited by 45
- 2018CODIT: Code Editing with Tree-Based Neural Modelscited by 40
- 2024REPOFUSE: Repository-Level Code Completion with Fused Dual Contextcited by 32
- 2023InstructCoder: Instruction Tuning Large Language Models for Code Editingcited by 29
- 2024Revisiting Code Similarity Evaluation with Abstract Syntax Tree Edit Distancecited by 26
- 2023Fixing Large Language Models' Specification Misunderstanding for Better Code Generationcited by 17
- 2023GrACE: Generation using Associated Code Editscited by 16
- 2024Does Your Neural Code Completion Model Use My Code? A Membership Inference Approachcited by 11
- 2022Overwatch: Learning Patterns in Code Edit Sequencescited by 8
- 2023Automated Code Editing with Search-Generate-Modifycited by 8
- 2022Automatically Recommend Code Updates: Are We There Yet?cited by 7
- 2019Structural Language Models of Codecited by 3
- 2024A Naive Approach for Automatic Line-level Code Completioncited by 2
Built on by
Citing papers per year: 2025: 2.
2 papers.