Re:LLM Powered Autonomous Agents
by vianneychevalie on Tuesday June 27, 2023 @05:29AM
(#36490732)
A chain of LLMs can work in that regard, using intermediary prompts that feed answers to the next prompt. Make the LLM build a list of sections, then make it fill them with examples, then make it enrich the text. Maybe a last layer for error correction, clarity, removing mentions of "as an AI model", etc.
[ Parent
| on HN ]
There are no comments on this comment yet.