Share your thoughts, 1 month free Claude Pro on usSee more

Code Debugging on HumanEval

96.3Accuracy

MGDebugger

Updated 3mo ago

Evaluation Results

Method	Links
MGDebugger 2024.10		96.3	-
MGDebugger 2024.10		95.7	-
Reflexion 2024.10		94.5	-
MGDebugger 2024.10		94.5	76.3
MGDebugger 2024.10		94.5	77.5
Reflexion 2024.10		93.9	-
Self-Debugging (Expl.) 2024.10		93.3	-
Self-Debugging (Expl.) 2024.10		93.3	-
MGDebugger 2024.10		92.7	-
MGDebugger 2024.10		91.5	64.1
Reflexion 2024.10		90.9	60.5
No-Debugging 2024.10		90.2	-
No-Debugging 2024.10		89.6	-
Reflexion 2024.10		89.6	-
Self-Debugging (Expl.) 2024.10		89.6	57.5
Simple Feedback 2024.10		88.4	52.5
Self-Debugging (Expl.) 2024.10		87.8	48.7
Reflexion 2024.10		87.8	48.7
Self-Debugging (Expl.) 2024.10		87.2	44.7
Self-Debugging (Expl.) 2024.10		86.6	-
Reflexion 2024.10		86.6	45
Self-Debugging (Trace) 2024.10		86	39.5
Self-Edit 2024.10		86	42.5
Simple Feedback 2024.10		85.4	38.5
Self-Debugging (Trace) 2024.10		84.8	35.9
LDB (Block) 2024.10		84.1	31.6
Self-Edit 2024.10		84.1	33.3
Self-Debugging (Trace) 2024.10		84.1	35
LDB (Block) 2024.10		83.5	32.5
LDB (Line) 2024.10		83.5	32.5
Self-Edit 2024.10		82.9	26.3
Simple Feedback 2024.10		82.3	23.7
LDB (Line) 2024.10		82.3	23.7
LDB (Function) 2024.10		82.3	27.5
LDB (Function) 2024.10		81.7	21.1
LDB (Function) 2024.10		80.5	17.9
LDB (Line) 2024.10		79.9	15.4
No-Debugging 2024.10		79.3	-
LDB (Block) 2024.10		79.3	12.8
No-Debugging 2024.10		76.8	-
No-Debugging 2024.10		76.2	-
No-Debugging 2024.10		75.6	-