2 posts found
You're not imagining it. Research tested 18 frontier models and every single one degrades as the conversation gets longer. Here's the science behind it.
Research proved that LLMs attend strongly to the beginning and end of context but poorly to the middle. This is a mathematical property, not a bug.