logoalt Hacker News

thesmtsolver2yesterday at 11:16 PM0 repliesview on HN

Did they though? That is the lore. You can’t really compare recommender system performance across different populations and products.

Unlike common benchmarks for LLMs.