If generic framework and domain specialist are both important, shouldn’t we train two stream of models (a generalist, and a batch of specialist), then report the performance?
What if evaluation metrics on LLM is designed on top of agentic workflow?! Would LLM surpass human in most intelligence work? 🫣
United States Trends
- 1. #FanCashDropPromotion 1,142 posts
- 2. Summer Walker 23.4K posts
- 3. Good Friday 53.6K posts
- 4. #NXXT_Earnings N/A
- 5. #FridayVibes 4,530 posts
- 6. Wale 37.1K posts
- 7. #GringosVenezuelaNoSeToca 2,924 posts
- 8. Happy Friyay 1,359 posts
- 9. #FinallyOverIt 7,693 posts
- 10. Saylor 37.7K posts
- 11. Meek 6,359 posts
- 12. Go Girl 25.6K posts
- 13. Doug McMillon N/A
- 14. $BTC 121K posts
- 15. Red Friday 3,367 posts
- 16. Robbed You 4,972 posts
- 17. Bill Clinton 57.1K posts
- 18. Monaleo 2,360 posts
- 19. Bubba 10.2K posts
- 20. SONIC RACING 2,742 posts
Loading...
Something went wrong.
Something went wrong.