Ai2 @allen_ai · 26 Aug 2026
A Thai research team adapted our Dolma data-curation toolkit to build Mangosteen, a 47B-token corpus for Thai LLMs. They used Dolma to filter widely used web datasets into a smaller corpus that improved Thai LLM performance despite using less data. 🧵 https://t.co/exvf02cy4U https://t.co/b3Ca8Xipyw
6 809Views
33Likes
6Reposts
6Replies
5Quotes
17Bookmarks
Is that a lot?
10.6×vs this author's median645 views is typical
83Percentile for this authorof 18 recent posts
2.29×vs 10K–100K median2 968 views is typical
7.91%Reachviews ÷ followers
0.73%Engagement rateof viewers reacted