This repo contains a CSV file containing about 2000 articles from the Straits Times from 04 Jan 2022 to 07 Nov 2023.
The file was generated using a script adapted from Ari's Scrapeyard: Straits Times.
A set of test queries and expected results are provided in test.csv
. Relevance information was assigned using boolean retrieval techniques, computed in the extract.ipynb
notebook, and saved to relevance.csv
.
A sample of predictions generated by lmsys/vicuna-7b-v1.5
for the test queries are provided in pred.csv
.