GPT-4o's Memory Breakthrough (Needle in a Needlestack)

Re:GPT-4o's Memory Breakthrough (Needle in a Needlestack) by m3kw9 on Tuesday May 14, 2024 @05:05PM (#40360164)

I thought google Gemini had almost perfect needle in haystack performance inside 1 million tokens?

[ Parent | on HN ]

Re:GPT-4o's Memory Breakthrough (Needle in a Needlestack) by sftombu on Tuesday May 14, 2024 @05:21PM (#40360298)

The reason I made Needle in a needlestack is the LLMs are getting to good at needle in a haystack. Until GPT-4o, no model was good at the NIAN benchmark.

[ Parent | on HN ]