OpenAI researchers reveal an algorithm by which LLMs can learn to better explain themselves to their users and improve the legibility of their outputs (Carl Franzen/VentureBeat)

Carl Franzen / VentureBeat:
OpenAI researchers reveal an algorithm by which LLMs can learn to better explain themselves to their users and improve the legibility of their outputs  —  One of the most interesting and useful slang terms to emerge from Reddit in my opinion is ELI5, from its subreddit of the same name …

from Techmeme https://ift.tt/OsqjcWR