EleutherAI @AiEleuther · 21 Aug 2026
What kinds of training data shape different model capabilities? A @GeorgiaTech team used our fully open model flow to trace Olmo’s performance on social/general reasoning and social-science/STEM knowledge tests back to the types of text it trained on. 🧵 https://t.co/qrPoTrdj2B https://t.co/pSQvYB5WaS
14 683Views
113Likes
20Reposts
3Replies
4Quotes
62Bookmarks
Is that a lot?
3.22×vs this author's median4 567 views is typical
86Percentile for this authorof 7 recent posts
5.70×vs 10K–100K median2 578 views is typical
51.09%Reachviews ÷ followers
0.95%Engagement rateof viewers reacted