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

ConvFinQA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Conversational Financial Question AnsweringConvFinQA (test)
Accuracy78.81
30
Financial ReasoningConvFinQA
Accuracy85.7
23
Financial Numerical ReasoningConvFinQA (dev)
Execution Accuracy85.67
13
Financial Question AnsweringConvFinQA (test)
Execution Accuracy78.46
9
Conversational Financial Question AnsweringConvFinQA
Accuracy85
9
Conversational Numerical Reasoning Question AnsweringConvFinQA v1 (dev)
Execution Accuracy0.7846
8
Conversational Numerical Reasoning Question AnsweringConvFinQA v1 (test)
Execution Accuracy89.44
7
Multi-step Reasoning over Code DependenciesConvFinQA hard
Accuracy (Multi-step Reasoning)74.67
6
Fact RetrievalConvFinQA (dev)
R@392.4
5
Conversational Question AnsweringConvFinQA (dev)
Exact Match Acc43.41
4
Showing 10 of 10 rows