Ask HN: How do looped transformers work with respect to context windows?
By simianwords · 2026-10-03 · 1 points · 0 comments
Posting here as I haven't found the answer anywhere. I know that GPT Astra uses looping instead of reasoning tokens. My specific question is this: reasoning tokens pollute the context and makes compaction kick in quickly. Does this also happen to looped transformers like Astra?
Open the full discussion on BetterNews