When LLMs Lag Behind: Knowledge Conflicts from Evolving APIs in Code Generation

TL;DR

This paper systematically studies how evolving APIs impact LLM code generation, revealing that LLMs often struggle with outdated knowledge despite documentation and reasoning strategies, highlighting the need for better benchmarks.

Contribution

It introduces a benchmark of real-world API updates and evaluates LLMs' ability to adapt, revealing persistent outdated patterns and the limited effectiveness of current strategies.

Findings

LLMs only generate 42.55% executable code without comprehensive documentation.

Structured documentation and larger models improve performance but do not fully solve executability issues.

Reasoning strategies like Self-Reflection improve executable rate by 11%.

Abstract

The rapid evolution of software libraries creates a significant challenge for Large Language Models (LLMs), whose static parametric knowledge often becomes stale post-training. While retrieval-augmented generation (RAG) is commonly used to provide up-to-date API specifications, "context-memory conflict" arises when external instructions contradict a model's internal parametric knowledge. This paper presents a systematic empirical study of LLM code generation under API evolution (e.g., API deprecation, API modification, and API addition), by constructing a benchmark of 270 real-world updates from eight Python libraries. We evaluate four LLM families of 11 models. Our results show that without comprehensive documentation, LLMs struggle to prioritize external context, averaging only 42.55% of generated code examples are executable in the target environment. While structured documentation and larger model scales improve LLMs' ability to update adoption, they do not fully resolve executability issues with a low 66.36% executable rate. In addition, reasoning-based strategies (e.g., Self-Reflection) significantly boost LLMs' performance with 11% improvement on executable rate. Our findings highlight the persistence of outdated patterns from LLMs, even when API update specifications are provided, and emphasize the need for evolution-aware benchmarks and techniques.