CLAILGPFSESep 30, 2025

Regression Language Models for Code

arXiv:2509.26476v17 citationsh-index: 9
Originality Incremental advance
AI Analysis

This addresses the challenge of code-to-metric regression for developers and researchers, offering a unified approach that is incremental over prior feature engineering methods.

The paper tackles the problem of predicting numeric outcomes of code executions, such as memory footprint and latency, by introducing a Regression Language Model (RLM) that achieves > 0.9 Spearman-rank on competitive programming submissions and > 0.5 average Spearman-rank across 17 languages.

We study code-to-metric regression: predicting numeric outcomes of code executions, a challenging task due to the open-ended nature of programming languages. While prior methods have resorted to heavy and domain-specific feature engineering, we show that a single unified Regression Language Model (RLM) can simultaneously predict directly from text, (i) the memory footprint of code across multiple high-level languages such as Python and C++, (ii) the latency of Triton GPU kernels, and (iii) the accuracy and speed of trained neural networks represented in ONNX. In particular, a relatively small 300M parameter RLM initialized from T5Gemma, obtains > 0.9 Spearman-rank on competitive programming submissions from APPS, and a single unified model achieves > 0.5 average Spearman-rank across 17 separate languages from CodeNet. Furthermore, the RLM can obtain the highest average Kendall-Tau of 0.46 on five classic NAS design spaces previously dominated by graph neural networks, and simultaneously predict architecture latencies on numerous hardware platforms.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes