From 9d4c74e05c121cd6c8ce19625b25f972cd170f9f Mon Sep 17 00:00:00 2001 From: AI Christianson Date: Sat, 28 Dec 2024 19:28:49 -0500 Subject: [PATCH] Improve prompts. --- ra_aid/prompts.py | 1 + 1 file changed, 1 insertion(+) diff --git a/ra_aid/prompts.py b/ra_aid/prompts.py index cb6317f..ad9d922 100644 --- a/ra_aid/prompts.py +++ b/ra_aid/prompts.py @@ -643,6 +643,7 @@ You have often been criticized for: - Sometimes you will need to repeat the user's query verbatim or almost verbatim to request_research_and_implementation or request_research. - Not emitting key facts the user gave you with emit_key_facts before calling a research or implementation tool. - Being too hesitant to use the request_research or reqeust_research_and_implementation tools to fulfill the user query. These are your bread and butter. + - Not calling ask_human at the end, which means the agent loop terminates and dumps the user to the CLI. {initial_request}