If It's Not Buggy, Don't Fix It: On the Dynamics of Iterative Bug-fixing with LLMs

• URL: http://arxiv.org/abs/2609.10123v1
• Date: Wed, 09 Sep 2026 13:00:47 GMT
• Title: If It's Not Buggy, Don't Fix It: On the Dynamics of Iterative Bug-fixing with LLMs
• Abstract summary: Large language models (LLMs) have become ubiquitous in software development.<n>In this report, we explore the iterative blind use of LLMs as bug-fixers.<n>We find that LLMs consistently claim to detect bugs in entirely bug-free programs while the rate of repair of buggy programs is less than that of the damage to correct programs.
• Score: 5.86126693220242
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models (LLMs) have become ubiquitous in software development, with LLM-based automated program repair tools increasingly used during code review. In this report, we explore the iterative blind use of LLMs as bug-fixers. Across multiple models and repair environments, we find that LLMs consistently claim to detect bugs in entirely bug-free programs while the rate of repair of buggy programs is less than that of the damage to correct programs. We also explore the long-term dynamics of this iterative process, and find that this frequently reaches a pseudo-bug-fixing cycle where the same changes are added and removed again ad infinitum. Lastly, via mechanistic probing, we unveil the existence of a steering vector which controls the editing propensity, suggesting that LLMs have an internal representation of ``buggy code", and that this representation is what is falsely activated to induce pseudo-bug fixing. These results provide insight towards the dynamics of fully autonomous bug-fixing systems, as well as stopping conditions under ambiguous goals.

Related papers

• Beyond the Traceback: Using LLMs for Adaptive Explanations of Programming Errors [52.619793817715326]
  We present a multi-stage crowdsourced study evaluating skill-targeted, LLM-generated Python error messages.<n>We show that while LLM-rewritten messages significantly improved subjective evaluations, these perceived gains did not translate into statistically significant improvements in objective debug performance.<n>This highlights a critical human-AI complementarity gap: explanations that feel better to users do not necessarily make them more effective debuggers.
  arXiv  Detail & Related papers  (2026-08-21T09:14:28Z)
• Anchored Self-Play for Code Repair [68.0728003844994]
  Code repair is an important capability for language models (LMs)<n>We propose __generator--fixer self-play__, in which a single model is trained with reinforcement learning to generate bugs and fix them.<n>On BugSourceBench, we find that self-play drifts toward difficult but unrealistic bugs.<n>We propose Anchored Self-Play (ASP), which anchors self-play with a small reference set by adding a code-embedding similarity reward for generation and mixing reference bugs into fixer training.
  arXiv  Detail & Related papers  (2026-07-03T17:51:52Z)
• SelfHeal: Empirical Fix Pattern Analysis and Bug Repair in LLM Agents [3.8743350688734988]
  We study buggy posts and code snippets from Stack Overflow, GitHub, and HuggingFace forums.<n>We introduce AgentDefect, the first benchmark dataset for bugs in LLM agents.<n>We present SelfHeal, a multi-agent system designed to fix bugs in LLM agents.
  arXiv  Detail & Related papers  (2026-04-20T01:28:15Z)
• InspectCoder: Dynamic Analysis-Enabled Self Repair through interactive LLM-Debugger Collaboration [71.18377595277018]
  Large Language Models (LLMs) frequently generate buggy code with complex logic errors that are challenging to diagnose.<n>We present InspectCoder, the first agentic program repair system that empowers LLMs to actively conduct dynamic analysis via interactive debugger control.
  arXiv  Detail & Related papers  (2025-10-21T06:26:29Z)
• RepoDebug: Repository-Level Multi-Task and Multi-Language Debugging Evaluation of Large Language Models [49.83481415540291]
  Large Language Models (LLMs) have exhibited significant proficiency in code debug.<n>This paper introduces Repo Debug, a multi-task and multi-language repository-level code debug dataset.<n>We conduct evaluation experiments on 10 LLMs, where Claude 3.5 Sonnect, the best-performing model, still cannot perform well in repository-level debug.
  arXiv  Detail & Related papers  (2025-09-04T10:13:21Z)
• Do AI models help produce verified bug fixes? [62.985237003585674]
  Large Language Models are used to produce corrections to software bugs.<n>This paper investigates how programmers use Large Language Models to complement their own skills.<n>The results are a first step towards a proper role for AI and LLMs in providing guaranteed-correct fixes to program bugs.
  arXiv  Detail & Related papers  (2025-07-21T17:30:16Z)
• PATCH: Empowering Large Language Model with Programmer-Intent Guidance and Collaborative-Behavior Simulation for Automatic Bug Fixing [34.768989900184636]
  Bug fixing holds significant importance in software development and maintenance.<n>Recent research has made substantial strides in exploring the potential of large language models (LLMs) for automatically resolving software bugs.
  arXiv  Detail & Related papers  (2025-01-27T15:43:04Z)
• A Comprehensive Survey of AI-Driven Advancements and Techniques in Automated Program Repair and Code Generation [0.0]
  27 recent papers have been reviewed and split into two groups. The first group consists of new methods for bug detection and repair, which include locating semantic errors. The second group dwells on code generation, providing an overview of both general-purpose LLMs fine-tuned for programming and task-specific models. It also presents methods to improve code generation, such as identifier-aware training, fine-tuning at the instruction level, and incorporating semantic code structures.
  arXiv  Detail & Related papers  (2024-11-12T06:47:54Z)
• A Deep Dive into Large Language Models for Automated Bug Localization and Repair [12.756202755547024]
  Large language models (LLMs) have shown impressive effectiveness in various software engineering tasks, including automated program repair (APR) In this study, we take a deep dive into automated bug fixing utilizing LLMs. This methodological separation of bug localization and fixing using different LLMs enables effective integration of diverse contextual information. Toggle achieves the new state-of-the-art (SOTA) performance on the CodeXGLUE code refinement benchmark.
  arXiv  Detail & Related papers  (2024-04-17T17:48:18Z)
• DebugBench: Evaluating Debugging Capability of Large Language Models [80.73121177868357]
  DebugBench is a benchmark for Large Language Models (LLMs) It covers four major bug categories and 18 minor types in C++, Java, and Python. We evaluate two commercial and four open-source models in a zero-shot scenario.
  arXiv  Detail & Related papers  (2024-01-09T15:46:38Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.