Datasets:

Modalities:
Text
Formats:
parquet
Languages:
English
Size:
< 1K
ArXiv:
Libraries:
Datasets
pandas
License:
natolambert commited on
Commit
6adeed3
1 Parent(s): d1ecb0f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +16 -0
README.md CHANGED
@@ -1,6 +1,22 @@
1
  ---
2
  license: apache-2.0
 
 
 
 
 
 
 
 
 
 
3
  ---
 
 
 
 
 
 
4
  To create the dataset, we do the following for our internal tooling.
5
  * rename `turns` to `prompts`,
6
  * add empty `reference` to remaining prompts (for HF Datasets),
 
1
  ---
2
  license: apache-2.0
3
+ task_categories:
4
+ - question-answering
5
+ - conversational
6
+ language:
7
+ - en
8
+ tags:
9
+ - evaluation
10
+ pretty_name: MT Bench
11
+ size_categories:
12
+ - n<1K
13
  ---
14
+
15
+ # MT Bench by LMSYS
16
+ This set of evaluation prompts is created by the [LMSYS org](https://huggingface.co/lmsys) for better evaluation of chat models.
17
+ For more information, see the [paper](https://arxiv.org/abs/2306.05685).
18
+
19
+ ### Dataset creation
20
  To create the dataset, we do the following for our internal tooling.
21
  * rename `turns` to `prompts`,
22
  * add empty `reference` to remaining prompts (for HF Datasets),