From Guessing to Placeholding: A Cost-Theoretic Framework for Uncertainty-Aware Code Completion

TL;DR

This paper introduces a cost-theoretic framework for uncertainty-aware code completion, enabling models to output placeholders at uncertain positions to reduce editing costs and improve user experience.

Contribution

It proposes Adaptive Placeholder Completion (APC), a novel approach that strategically outputs placeholders, backed by theoretical analysis and reinforcement learning training methods.

Findings

APC reduces expected editing costs by up to 50%.

Models trained with APC maintain standard code completion performance.

Theoretical analysis shows a critical entropy threshold for effective placeholder use.

Abstract

While Large Language Models (LLMs) have demonstrated exceptional proficiency in code completion, they typically adhere to a Hard Completion (HC) paradigm, compelling the generation of fully concrete code even amidst insufficient context. Our analysis of 3 million real-world interactions exposes the limitations of this strategy: 61% of the generated suggestions were either edited after acceptance or rejected despite exhibiting over 80% similarity to the user's subsequent code, suggesting that models frequently make erroneous predictions at specific token positions. Motivated by this observation, we propose Adaptive Placeholder Completion (APC), a collaborative framework that extends HC by strategically outputting explicit placeholders at high-entropy positions, allowing users to fill directly via IDE navigation. Theoretically, we formulate code completion as a cost-minimization problem under uncertainty. Premised on the observation that filling placeholders incurs lower cost than correcting errors, we prove the existence of a critical entropy threshold above which APC achieves strictly lower expected cost than HC. We instantiate this framework by constructing training data from filtered real-world edit logs and design a cost-based reward function for reinforcement learning. Extensive evaluations across 1.5B--14B parameter models demonstrate that APC reduces expected editing costs from 19% to 50% while preserving standard HC performance. Our work provides both a theoretical foundation and a practical training framework for uncertainty-aware code completion, demonstrating that adaptive abstention can be learned end-to-end without sacrificing conventional completion quality.