From 6cc93f48ec4c9c222bd8b1b94a98f2da6a0b7660 Mon Sep 17 00:00:00 2001 From: osmarks Date: Sat, 14 Sep 2024 06:17:41 +0000 Subject: [PATCH] =?UTF-8?q?Edit=20=E2=80=98autogollark=E2=80=99?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- autogollark.myco | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/autogollark.myco b/autogollark.myco index 0cfa34e..51d3746 100644 --- a/autogollark.myco +++ b/autogollark.myco @@ -1 +1,7 @@ -Autogollark is an [[emulation]] or primitive [[beta upload]] of [[gollark]] using a proprietary dataset of dumped [[Discord]] messages, [[semantic search]] and [[in-context learning]] on a [[base model]]. Currently, it uses LLaMA-3.1-405B in FP8, [[AutoBotRobot]] as a frontend and a custom [[PGVector]]-based search API. While not consistently coherent, Autogollark is generally able to approximately match personality and typing style parameters. \ No newline at end of file +Autogollark is an [[emulation]] or primitive [[beta upload]] of [[gollark]] using a proprietary dataset of dumped [[Discord]] messages, [[semantic search]] and [[in-context learning]] on a [[base model]]. Currently, it uses LLaMA-3.1-405B in FP8 via Hyperbolic, [[AutoBotRobot]] as a frontend and a custom [[PGVector]]-based search API. While not consistently coherent, Autogollark is able to approximately match personality and typing style. + +TODO: + +* reformat dataset to include longer-form conversation chunks for increased long-term coherence +* fix emoji/ping formatting +* writeable memory? \ No newline at end of file