Hacker News new | ask | show | jobs
by riku_iki 836 days ago
LLMs already have short term memory: context window when they predict next token?