Papers
arxiv:2606.18286

CODEBLOCK: Learning to Supervise Code at the Right Granularity

Published on Jun 10
Authors:
,
,
,
,
,
,

Abstract

CodeBlock is a structure-aware sparse supervision framework for code language model fine-tuning that selects syntactically coherent code blocks instead of individual tokens to improve training efficiency and performance.

Supervised fine-tuning of code LLMs typically applies uniform cross-entropy loss to all response tokens, implicitly assuming that every token provides equally useful learning signal. Recent token-level selection methods challenge this assumption in natural-language SFT by supervising only high-value tokens. However, directly transferring token-level masking to code can break syntactically and semantically coherent program units, because code depends on structural completeness and definition-use relations. We therefore propose CodeBlock, a structure-aware sparse supervision framework that selects structure-complete code evidence rather than isolated tokens. CodeBlock first selects high-quality instruction-response pairs, then partitions code responses into syntactically coherent coding items, estimates their utility by aggregating generalized cross-entropy over core logic tokens, and reranks them with data-flow reach and bridge signals to prioritize blocks that propagate or connect important program dependencies. During training, the full response remains available as context, while loss is applied only to selected code items and informative natural-language tokens. Experiments on six code-generation benchmarks show that CodeBlock achieves stronger average pass@1 than full-token SFT and competitive selection baselines, while using only 1.9% of supervised response tokens.

Community

Sign up or log in to comment

Get this paper in your agent:

hf papers read 2606.18286
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2606.18286 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2606.18286 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2606.18286 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.