How Should Diffusion Language Models Edit Code?
cs.SE, cs.AI
Submitted: 2026-09-29
Updated: 2026-09-29
Code: https://github.com/inclusionAI/LLaDA2.X
Sources
- Efficient Training of Language Models to Fill in the Middle
- Can It Edit? Evaluating the Ability of Large Language Models to Follow Code Editing Instructions
- Evaluating Large Language Models Trained on Code
- Stable-DiffCoder: Pushing the Frontier of Code Diffusion Large Language Model
- DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation
- Levenshtein Transformer
- Edit Flows: Flow Matching with Edit Operations
- SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
- Any-Order Flexible Length Masked Diffusion
- StarCoder 2 and The Stack v2: The Next Generation
- Large Language Diffusion Models
- Improved Large Language Diffusion Models
- DiffusER: Discrete Diffusion via Edit-based Reconstruction
- Simple and Effective Masked Diffusion Language Models
- Simplified and Generalized Masked Diffusion for Discrete Data
- DreamOn: Diffusion Language Models For Code Infilling Beyond Fixed-size Canvas
- Dream-Coder 7B: An Open Diffusion Language Model for Code
- Predict, Don't Iterate: Efficient Adaptive-Length Infilling for Diffusion Language Models
- Dream 7B: Diffusion Large Language Models
- Bridging the Editing Gap in LLMs: FineEdit for Precise and Targeted Text Modifications
Related papers
- Falsification-Based Verification of LLM-Generated Optimization Models: Sound Test Batteries and Their Detection Limits
- GitSkills: A Dataset of Agent Skills on GitHub
- SABER: Benchmarking Operational Safety of LLM Coding Agents in Stateful Project Workspaces
- PackMonitor: Enabling Zero Package Hallucinations Through Decoding-Time Monitoring
- IntentCoding: Amplifying User Intent in Code Generation
- Incentives and Outcomes in Bug Bounties