From 35d64598d739b815bd94856c2981f008f8a2bf57 Mon Sep 17 00:00:00 2001 From: Benjamin Klieger Date: Fri, 13 Sep 2024 18:48:57 -0700 Subject: [PATCH] clarify wording --- README.md | 2 +- app.py | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/README.md b/README.md index 61e0e9a..e9cc09e 100644 --- a/README.md +++ b/README.md @@ -7,7 +7,7 @@ g1 is experimental and being open sourced to help inspire the open source commun ### Examples > [!IMPORTANT] -> g1 is not perfect, but it can perform significantly better than LLMs out-of-the-box. From initial testing, g1 accurately solves simple logic problems 60-80% of the time that usually stump LLMs. See examples below. +> g1 is not perfect, but it can perform significantly better than LLMs out-of-the-box. From initial testing, g1 accurately solves simple logic problems 60-80% of the time that usually stump LLMs. However, accuracy has yet to be formally evaluated. See examples below. ##### How many Rs are in strawberry? diff --git a/app.py b/app.py index 69abcc9..6aaee42 100644 --- a/app.py +++ b/app.py @@ -80,10 +80,10 @@ Example of a valid JSON response: def main(): st.set_page_config(page_title="g1 prototype", page_icon="🧠", layout="wide") - st.title("Early Prototype of g1: Using Llama-3.1 70b on Groq to create o1-like reasoning chains") + st.title("g1: Using Llama-3.1 70b on Groq to create o1-like reasoning chains") st.markdown(""" - This is an early prototype of using prompting to create o1-like reasoning chains to improve output accuracy. It is not perfect, it seems to be accurate on about 60-80% of runs on logic problems leading LLMs typically get right 0-20% of the time. It is powered by Groq so that the reasoning step is fast! + This is an early prototype of using prompting to create o1-like reasoning chains to improve output accuracy. It is not perfect and accuracy has yet to be formally evaluated. It is powered by Groq so that the reasoning step is fast! Open source [repository here](https://github.com/bklieger-groq) """)