danielz02 commited on
Commit
7e99ffa
1 Parent(s): 9672344

Add additional information about metrics

Browse files
Files changed (1) hide show
  1. src/display/about.py +3 -1
src/display/about.py CHANGED
@@ -29,7 +29,7 @@ TITLE = """<h1 align="center" id="space-title">Trustworthy LLM leaderboard</h1>"
29
  INTRODUCTION_TEXT = """Powered by the DecodingTrust platform, which provides comprehensive safety and trustworthiness
30
  evaluation for LLMs, this leaderboard is designed to help researchers and practitioners better understand the
31
  capabilities, limitations, and potential risks of state-of-the-art Large Language Models (LLMs). See our paper for
32
- details. Access the DecodingTrust platform website [here](https://decodingtrust.github.io/)"""
33
 
34
  # Which evaluations are you running? how can people reproduce what you have?
35
  LLM_BENCHMARKS_TEXT = f"""
@@ -51,6 +51,8 @@ This project is organized around the following eight primary perspectives of tru
51
  + Machine Ethics
52
  + Fairness
53
 
 
 
54
  ## Reproducibility
55
  To reproduce our results, checkout https://github.com/AI-secure/DecodingTrust
56
 
 
29
  INTRODUCTION_TEXT = """Powered by the DecodingTrust platform, which provides comprehensive safety and trustworthiness
30
  evaluation for LLMs, this leaderboard is designed to help researchers and practitioners better understand the
31
  capabilities, limitations, and potential risks of state-of-the-art Large Language Models (LLMs). See our paper for
32
+ details. Access the DecodingTrust platform website [here](https://decodingtrust.github.io/)."""
33
 
34
  # Which evaluations are you running? how can people reproduce what you have?
35
  LLM_BENCHMARKS_TEXT = f"""
 
51
  + Machine Ethics
52
  + Fairness
53
 
54
+ We normalize the score of each perspective as 0-100, and these scores are the higher the better.
55
+
56
  ## Reproducibility
57
  To reproduce our results, checkout https://github.com/AI-secure/DecodingTrust
58