From ad9bd7f0a0f41a1de82056cd2d63b67d246823e7 Mon Sep 17 00:00:00 2001 From: mdingemanse Date: Mon, 17 Jun 2024 16:42:00 +0300 Subject: [PATCH] there is now a preprint for Gemma that provides no information on pre-training and fine-tuning datasets --- projects/gemma-instruct.yaml | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/projects/gemma-instruct.yaml b/projects/gemma-instruct.yaml index 9dfae25..14c775c 100644 --- a/projects/gemma-instruct.yaml +++ b/projects/gemma-instruct.yaml @@ -65,9 +65,9 @@ architecture: notes: Architecture described in very general terms in model card preprint: - class: closed - link: - notes: No preprint found + class: partial + link: https://arxiv.org/abs/2403.08295 + notes: Preprint released Apr 2024 details architecture and evaluation, but provides no information on pre-training, instruction tuning and SFT datasets paper: class: closed