AI Paper Atlas

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

Incoherence as Oracle-less Measure of Error in LLM-Based Code Generation

Thomas Jean-Michel Valentin, Ardi Madadi, Gaetano Sapia, Marcel Böhme

AAAI 2026 · Large language models › Agents, code and math · approach: Execution-based code verification · 16 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
sample more programs per task (m from 10 to 50) (ablation)Incoherence detection (GPT-4o) · MBPPdetection rate: 0.7243 -> 0.8105Improved
sample more programs per task (m from 25 to 50) (ablation)Incoherence detection (GPT-4o) · HumanEvaldetection rate: 0.8267 -> 0.8182No clear effect
sample a second program (m=2 vs m=1) (ablation)Incoherence detection (GPT-4o) · MBPPdetection rate: 0.0000 -> 0.3974Improved
raise sampling temperature from 0.2 to 1.0 (ablation)Incoherence detection (GPT-4o) · MBPPdetection rate: 0.5422 -> 0.8050Improved
raise sampling temperature from 0.2 to 1.0 (ablation)Incoherence detection (GPT-4o) · HumanEvaldetection rate: 0.5556 -> 0.7600Improved
generate more test inputs (n from 100 to 10000) (ablation)Incoherence detection (GPT-4o) · MBPPdetection rate: 0.6811 -> 0.7445Improved
generate more test inputs (n from 1000 to 10000) (ablation)Incoherence detection (GPT-4o) · HumanEvaldetection rate: 0.7286 -> 0.7222No clear effect
use zero incoherence as a filter for likely-correct programs16 LLMs · MBPPmean error: 0.3009 -> 0.1866Improved
use zero incoherence as a filter for likely-correct programs16 LLMs · HumanEvalmean error: 0.1050 -> 0.0471Improved
rank LLMs by incoherence instead of by pass@116 LLMs · HumanEvalSpearman rho of rankings: rho = 0.94Improved

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