AINativeBench / data /processed /RQ1 /MarkdownValidator-MCP /evaluation_results.csv
王子睿
restructure + add files
8c10cf2
Raw
History Blame Contribute Delete
768 Bytes
model,num_samples,exact_match,in_order_match,any_order_match,precision,recall,single_tool_use,unique_path_ratio,path_entropy
GPT-5,60,1.0,1.0,1.0,1.0,1.0,1.0,0.016666666666666666,0.0
GPT-4o-mini,60,0.9833333333333333,1.0,1.0,0.9958333333333333,1.0,1.0,0.03333333333333333,0.1222915970693747
DeepSeek-V3-1,60,1.0,1.0,1.0,1.0,1.0,1.0,0.016666666666666666,0.0
DeepSeek-R1,60,0.9666666666666667,1.0,1.0,0.9916666666666667,1.0,1.0,0.03333333333333333,0.21084230031853216
Gemini-2.5-flash,60,0.2,0.23333333333333334,0.25,0.8839586339586342,0.8638888888888892,0.25,0.21666666666666667,0.7315687590573717
Gemini-2.5-flash-nothinking,60,0.1,0.1,0.1,0.9537037037037038,0.7277777777777777,0.1,0.1,0.5411669865395448
Qwen3-235b,60,1.0,1.0,1.0,1.0,1.0,1.0,0.016666666666666666,0.0