Wouldn’t you expect this to be affected pretty heavily by availability bias? Meaning a large determinant of the data is going to be if the articles available to the LLM for a particular time period are for or against a particular President on the proposed topics.
It might be different if you were searching a comprehensive system, like 100% of court records, for incidents, but it sounds like your search would be of news articles, web sites, and similar, which is going to be heavily influenced by what reporters found desirable to report on at the time. It’s a dataset which will highlight unusual events and incidents damaging to political opponents over systemic reporting of events from a neutral observer.
Wouldn’t you expect this to be affected pretty heavily by availability bias? Meaning a large determinant of the data is going to be if the articles available to the LLM for a particular time period are for or against a particular President on the proposed topics.
It might be different if you were searching a comprehensive system, like 100% of court records, for incidents, but it sounds like your search would be of news articles, web sites, and similar, which is going to be heavily influenced by what reporters found desirable to report on at the time. It’s a dataset which will highlight unusual events and incidents damaging to political opponents over systemic reporting of events from a neutral observer.