如何让包含超过 100 万个连续步骤的 AI 任务,不会出现任何错误? 用便宜模型对每一个步骤进行投票。 让我马上就想到了是不是可以用在AI炒股。
I just read a paper that completely broke my brain. It describes a system that solved an AI task with over 1,000,000 sequential steps... with ZERO errors. Using AI models that are known to be flaky and make mistakes. How is that even possible? 🤯 We all know LLMs have an…
169
550
3千
3千
300千
0
0
0
0
79
United States 趋势
- 1. Mets 36.9K posts
- 2. Dodgers 42.7K posts
- 3. Diaz 78.1K posts
- 4. Maresca 38.7K posts
- 5. Kentucky State University 3,165 posts
- 6. White Sox 3,351 posts
- 7. 2026 MLB Draft 2,519 posts
- 8. Stearns 12.1K posts
- 9. Slept 20.5K posts
- 10. Roch Cholowsky N/A
- 11. GeForce Season 4,382 posts
- 12. Paolo 9,863 posts
- 13. Dictionary 8,329 posts
- 14. Rivers 83.4K posts
- 15. Soto 15.5K posts
- 16. #NBACup N/A
- 17. Alonso 62.6K posts
- 18. #LightningStrikes N/A
- 19. UMBC N/A
- 20. #TrumpInflationCrisis 12K posts
Loading...
Something went wrong.
Something went wrong.