search

jbreffle / acx-prediction-contest

Analysis of the ACX 2023 Prediction Contest with an interactive Streamlit app
https://jbreffle.github.io/acx-app
5 stars 0 forks source link

Typo in outlier questions #6

Open jacobwood27 opened 4 months ago

jacobwood27 commented 4 months ago

Hi, thanks for putting this together! It looks like you pasted in the same question in this section:

Most questions had a high resposne rate. Two questions are slight outliers, with low response rates. They are:

Question 41. Will an image model win Scott Alexander’s bet on compositionality, to Edwin Chen’s satisfaction, in 2023? Which had a response rate of 75.99% Question 41. Will an image model win Scott Alexander’s bet on compositionality, to Edwin Chen’s satisfaction, in 2023? Which had a response rate of 76.93%

jbreffle commented 3 months ago

Those are actually dynamically calculated directly from the pandas dataframe. This seems to be an intermittent bug that might be hard to track the source of, so I've just replaced the dynamic computation with hard-coded text. Thanks!