AI Paper Atlas

Research map / Large language models / Agents, code and math

Towards Functional Correctness of Large Code Models with Selective Generation

Jaewoo Jeong, Taesoo Kim, Sangdon Park

ICML 2026 · Large language models › Agents, code and math · approach: Execution-based code verification · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
apply SCG selective generation on top of CodeTCodeT (GPT-3.5-Turbo) · MBPP-fFDR-CE: 0.491±0.042 -> 0.145±0.052Improved
apply SCG on top of LDBLDB (GPT-3.5-Turbo) · MBPP-fFDR-CE: 0.442±0.044 -> 0.142±0.043Improved
apply SCG on top of SFSSFS (GPT-3.5-Turbo) · HumanEval-fFDR-CE: 0.288±0.072 -> 0.158±0.058Improved
add selective generation (SCG) to GPT-4oGPT-4o · APPS-fFDR-CE: 0.431±0.011 -> 0.224±0.018Improved
use fuzzing-generated unit tests instead of 21 APPS tests (SCG vs SCG-small) (ablation)SCG (GPT-4o) · APPS-fselection efficiency: 0.339±0.012 -> 0.337±0.011No clear effect
include false entailment rate term eps_E (SCG vs SCG-H) (ablation)SCG (GPT-4o) · APPS-fFDR-CE: 0.280±0.020 -> 0.224±0.018Improved
include false entailment rate term eps_E (SCG vs SCG-H) (ablation)SCG (GPT-4o) · HumanEval-fFDR-CE: 0.145±0.123 -> 0.049±0.111Improved
use calibrated threshold instead of top-k% manual selection (SCG vs SCG-manual) (ablation)SCG (GPT-4o) · APPS-fFDR-CE: 0.291±0.014 -> 0.224±0.018Improved
use exact-match selective predictor (SCG-EM) rather than unit-test entailment (ablation)SCG-EM (GPT-4o) · APPS-fselection efficiency: 0.000±0.000Hurt
use verbalized confidence scoring function instead of normalized log-prob (ablation)SCG (GPT-4o) · APPS-fFDR-CE controlHurt
use smaller eps_E (more unit tests via fuzzing) (ablation)SCG (GPT-4o) · APPS-fFDR-CE bound tightnessImproved
increase number of fuzzed unit tests used for evaluation (ablation)FuzzEval · APPS-fFDR-CEImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy