ResearchTasksSentence-level Error DetectionSentence-level Error DetectionFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedDeltaBench CoT Diagnosis 1.0 (test)GPT-5 (BIG-Bench Prompt)43.2Precision3Mar 24, 2026HQ2A 1.0 (test)GPT-3.5-Turbo25.49Exact Accuracy1Feb 26, 2026