Update space
Browse files- app.py +1 -0
- src/about.py +4 -1
app.py
CHANGED
@@ -139,6 +139,7 @@ with demo:
|
|
139 |
with gr.TabItem("📝 About", elem_id="llm-benchmark-tab-table", id=5):
|
140 |
gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text")
|
141 |
|
|
|
142 |
'''
|
143 |
with gr.TabItem("🚀 Submit here! ", elem_id="llm-benchmark-tab-table", id=3):
|
144 |
with gr.Column():
|
|
|
139 |
with gr.TabItem("📝 About", elem_id="llm-benchmark-tab-table", id=5):
|
140 |
gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text")
|
141 |
|
142 |
+
|
143 |
'''
|
144 |
with gr.TabItem("🚀 Submit here! ", elem_id="llm-benchmark-tab-table", id=3):
|
145 |
with gr.Column():
|
src/about.py
CHANGED
@@ -25,7 +25,10 @@ TITLE = """<h1 align="center" id="space-title">Decentralized Arena</h1>"""
|
|
25 |
|
26 |
# What does your leaderboard evaluate?
|
27 |
INTRODUCTION_TEXT = """
|
28 |
-
|
|
|
|
|
|
|
29 |
"""
|
30 |
|
31 |
# Which evaluations are you running? how can people reproduce what you have?
|
|
|
25 |
|
26 |
# What does your leaderboard evaluate?
|
27 |
INTRODUCTION_TEXT = """
|
28 |
+
TL;DR: We release Decentralized Arena that automates and scales “Chatbot Arena” for LLM evaluation across various
|
29 |
+
fine-grained dimensions (e.g., math – algebra, geometry, probability; logical reasoning, social reasoning,
|
30 |
+
biology, chemistry, …). The evaluation is decentralized and democratic, with all LLMs participating
|
31 |
+
in evaluating others. It achieves a 97\% correlation with Chatbot Arena's overall rankings, while being fully transparent and reproducible.
|
32 |
"""
|
33 |
|
34 |
# Which evaluations are you running? how can people reproduce what you have?
|