Data Discovery and Query Solving on KramaBench Oracle Setting 2025
41.67Archaeology ScoreData Interpreter
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Data InterpreterFramework=Data Interpreter, Model=Gemini-2.5-Pro2025.09 | 41.67 | 13.29 | 28.39 | 13.09 | 32.97 | 63.13 | 33.57 | |
| DA-AgentFramework=DA-Agent, Model=Gemini-2.5-Pro2025.09 | 41.67 | 15.77 | 44.26 | 44.55 | 56.42 | 65.92 | 48.61 | |
| ReActFramework=ReAct, Model=Gemini-2.5-Pro2025.09 | 25 | 6.36 | 3.69 | 30.7 | 46.93 | 51.69 | 33.82 | |
| AutoGenFramework=AutoGen, Model=Gemini-2.5-Pro2025.09 | 25 | 5.29 | 26.03 | 23.46 | 39.28 | 49.41 | 31.77 | |
| DS-STARFramework=DS-STAR, Model=Gemini-2.5-Pro2025.09 | 25 | 19.08 | 55.24 | 58.8 | 59.66 | 70.18 | 52.55 |