4 ms·
That sounds incredibly disappointing given how high their benchmarks are, indicating they might be overtuned for those, similar to Llama4.
by jjani 2y ago
That sounds incredibly disappointing given how high their benchmarks are, indicating they might be overtuned for those, similar to Llama4.
- XCSme 2y agoYeah, I think so too. They seemed to be better at specific tasks, but worse overall, at broader tasks.