List your company

Hand-eye Calibration on Hand-eye calibration task

96.67Code Gen Success RateMCS + Debugger w/ Probing
3.0727.3751.6775.97Jun 2, 2026

Evaluation results

MethodLinks
MCS + Debugger w/ ProbingLLM Backbone=GPT-4o, Code Generation Strategy=MCS, Feedback Strategy=Debugger, Probing=true2026.0696.6786.670.0480.0110.170.012
CaP with Full APICode Generation Strategy=CaP, API Configuration=Full API2026.0690———0.0380.039
MCS + Debugger w/ ProbingLLM Backbone=Gemini, Code Generation Strategy=MCS, Feedback Strategy=Debugger, Probing=true2026.069073.330.0640.0640.0910.065
MCS + DebuggerLLM Backbone=GPT-4o, Code Generation Strategy=MCS, Feedback Strategy=Debugger, Probing=false2026.0673.33600.1740.0190.0740.014
MCS + Debugger w/ ProbingLLM Backbone=GPT-4.1 mini, Code Generation Strategy=MCS, Feedback Strategy=Debugger, Probing=true2026.0673.33400.0850.0680.3370.068
ProgPromptCode Generation Strategy=ProgPrompt2026.0666.67———0.0260.026
MCSLLM Backbone=GPT-4o, Code Generation Strategy=MCS2026.0623.336.670.0290.0290.3080.283
SPG + Debugger w/ ProbingLLM Backbone=GPT-4o, Code Generation Strategy=SPG, Feedback Strategy=Debugger, Probing=true2026.0610———0.4850.062
SPG + DebuggerLLM Backbone=GPT-4o, Code Generation Strategy=SPG, Feedback Strategy=Debugger, Probing=false2026.066.67———0.8020.802