3 ms·
Also, even if you want to be honest, at this point, probably every public or semipublic benchmark is part of CommonCrawl.
by antupis 1y ago
Also, even if you want to be honest, at this point, probably every public or semipublic benchmark is part of CommonCrawl.
- NitpickLawyer 1y agoTrue. And it's even worse than that, because each test probably gets "talked about" a lot in various places. And people come up with variants. And those variants get ingested. And then the whole thing becomes a mess. This was noticeable with the early Phi models. They were originally trained fully on synthetic data (cool experiment tbh) but the downside was that GPT3 / 4 was "distilling" benchmarks "hacks" into it. It became aparent when new benchmarks were released, after the published date, and there was one that measured "contamination" of about 20+%. Just from distillation.