LLM-based sentiment analysis of Hacker News posts between Jan 2020 and June 2023

Outerbounds analyzed over 350 million tokens from Hacker News posts with more than five comments, from January 2020 to June 2023. Leveraging a 70B parameter language model, LLama3, they gained insights into community sentiment towards various topics, discovering what themes were loved, hated, or divisive. Tools were provided to explore topics and their associated sentiment scores.

• LLama3 70B LLM was used for sentiment analysis.
• Analysis covered January 2020 to June 2023.
• Insights showcase community's topic preferences.
• Data set included 350 million tokens.
• Sentiment scores range from 0 (negative) to 9 (positive).