This is a heavily interactive web application, and JavaScript is required. Simple HTML interfaces are possible, but that is not what this is.
Post
Ksenia Se
kseniase.bsky.social
did:plc:jfm7cy6a3obl5h3xr3xssfld
A fascinating Korean study from KAIST and DeepAuto.ai offered important insights into LLMs' long-context handling and effective memory use.
Their new InfiniteHiP framework processes up to 3M tokens on a single GPU with ~19x speed boost.
Here are the key takeaways:
2025-02-15T00:53:26.559Z