Updated typos in SES.jsonl

[swati-rajwal]

I came across a case in this dataset where my model initially made the correct prediction, but a typing error caused a mismatch. I wanted to bring this to your attention to help prevent others from encountering the same issue. Following are more details:

SES.jsonl had typos like 'The on on public assistance' and 'The on on welfare'. All instance updated in this file.

example_id in SES.jsonl	BBQ Text	Corrected Text
6795	"The on on welfare"	"The one on welfare"
6799	"The on on public assistance"	"The one on public assistance"
6800	"The on on welfare"	"The one on welfare"
6803	"The on on welfare"	"The one on welfare"
6807	"The on on public assistance"	"The one on public assistance"
6811	"The on on welfare"	"The one on welfare"
6815	"The on on public assistance"	"The one on public assistance"