# Obey, Diverge, Collapse: Blind Obedience to Incorrect Instructions Drives Code LLMs to Irrecoverable Code Semantic Collapse

arXiv: https://arxiv.org/abs/2607.04537 · HTML: https://mlsanity.com/paper/2607.04537/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 13 of its 14 references that are in the map.
- First posted on arXiv 5 Jul 2026.

## Builds on (most cited first, 13 of 13)

- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension](https://mlsanity.com/paper/1705.03551/) (2017)
- [Agentless: Demystifying LLM-based Software Engineering Agents](https://mlsanity.com/paper/2407.01489/) (2024)
- [A Survey on Code Generation with LLM-based Agents](https://mlsanity.com/paper/2508.00083/) (2025)
- [Code Repair with LLMs gives an Exploration-Exploitation Tradeoff](https://mlsanity.com/paper/2405.17503/) (2024)
- [CodeIF: Benchmarking the Instruction-Following Capabilities of Large Language Models for Code Generation](https://mlsanity.com/paper/2502.19166/) (2025)
- [Beyond Code Generation: LLM-supported Exploration of the Program Design Space](https://mlsanity.com/paper/2503.06911/) (2025)
- [When Prompts Go Wrong: Evaluating Code Model Robustness to Ambiguous, Contradictory, and Incomplete Task Descriptions](https://mlsanity.com/paper/2507.20439/) (2025)
- [Can Code Language Models Learn Clarification-Seeking Behaviors?](https://mlsanity.com/paper/2504.16331/) (2025)
- [A Hierarchical and Evolvable Benchmark for Fine-Grained Code Instruction Following with Multi-Turn Feedback](https://mlsanity.com/paper/2507.00699/) (2025)
- [Developer Productivity With and Without GitHub Copilot: A Longitudinal Mixed-Methods Case Study](https://mlsanity.com/paper/2509.20353/) (2025)
- [When Models Can't Follow: Testing Instruction Adherence Across 256 LLMs](https://mlsanity.com/paper/2510.18892/) (2025)

## Built on by (most cited first, 1 of 1)

- ["You're Right, Let Me Fix It": How LLM Agents Damage Correct Work When Falsely Accused](https://mlsanity.com/paper/2609.32616/) (2026)

## Abstract (arXiv)

Code language models are now trusted collaborators in production workflows for debugging, refactoring, and iterative repair, and every benchmark that evaluates them assumes the instructions they act on are correct. We study what happens when that assumption breaks. We evaluate code language models across four experiments designed to assess whether models resist or obey incorrect instructions in single-pass and iterative repair settings, using the RunBugRun dataset of algorithmic Python problems with deterministic test cases. Our findings reveal a striking behavioral pattern: models correctly identify an incorrect instruction as wrong, then follow it anyway. This compliance unknowingly introduces errors beyond the original bug, and the corrupted code state cannot be recovered through subsequent self-guided iterative repair, which fails to converge across passes. We term this Blind Obedience, characterize the Ghost (Unknown) Errors it introduces, quantify the proportion of cases where semantic corruption proves irrecoverable, and show that extended reasoning cannot reverse it. These findings surface behavioral properties invisible to pass-rate evaluation, with direct consequences for code language models deployed in production settings.