Issue Genie Benchmark: Different responses in UI and Benchmark

[maze2498]

Hello, I am trying to add a benchmark dataset for my genie space.

When I ask the a question on the Genie space UI directly, I get the right output. However when I add the same question in the genie benchmark, the result is quite bad and the sql it uses in benchmark is incomplete.

[emma_s]

Hi, when you say the sql it generates is quite bad and missing, do you mean when you run the benchmark? The benchmark purposefully doesn't have any conversation history unlike the Genie Space. So sometimes the results can vary. Ie if you've asked a lot of questions before the one in your Genie Space there will be additional context th