Did you read the blog post? They compare to deepswe and call it out as the worst one for false positives (failed, but the benchmark assessed it as correct). It also has less language variance.
replyI mean yes that is what you'd say if you were writing a blog post about your new benchmark.
replySure, but they at least quantified it with data. It's not like they just dropped a sentence saying the above, they showed numbers.
reply