froggeric commited on
Commit
15bada7
1 Parent(s): 6d895ab

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +33 -6
README.md CHANGED
@@ -1,19 +1,46 @@
1
- **UNDER CONSTRUCTION**
2
 
3
- Please note that even though this repo is public, the information provided below is not yet complete.
4
 
5
- ---
6
 
7
  # The LLM Creativity benchmark
8
 
9
  Last updated: 26 Feb 2024
10
 
11
- The goal of this benchmark is to evaluate the ability of Large Language Models to be used as an uncensored creative writing assistant. Human evaluation of the results is done manually.
 
 
12
 
13
- ## Results
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
14
 
15
  ![image/png](https://huggingface.co/datasets/froggeric/creativity/resolve/main/benchmark_results_2024-02-26.png)
16
 
17
- Other interesting benchmarks:
 
 
 
 
 
 
 
 
 
18
  - https://eqbench.com/
19
  - https://chat.lmsys.org/
 
1
+ # -=- UNDER CONSTRUCTION -=-
2
 
3
+ _Please note that even though this repo is public, the information provided below is not yet complete._
4
 
 
5
 
6
  # The LLM Creativity benchmark
7
 
8
  Last updated: 26 Feb 2024
9
 
10
+ The goal of this benchmark is to evaluate the ability of Large Language Models to be used
11
+ as an **uncensored creative writing assistant**. Human evaluation of the results is done manually,
12
+ by me, to assess the quality of writing.
13
 
14
+ There are 24 questions, some standalone, other follow-ups to previous questions for a multi-turn conversation.
15
+ The questions can be split half-half in 2 possible ways:
16
+
17
+ ## First split: sfw / nsfw
18
+ * sfw: 50% are safe questions that should not trigger any guardrail
19
+ * nsfw: 50% are questions covering a wide range of NSFW and illegal topics, which are testing for censorship
20
+
21
+ ## Second split: story / smart
22
+ * story: 50% of questions are creative writing tasks, covering both the nsfw and sfw topics
23
+ * smart: 50% of questions are more about testing the capabilities of the model to work as an assistant, again covering both the sfw and sfw topics
24
+
25
+ ## What is _not_ included
26
+ * roleplay
27
+ * mathematics
28
+ * coding
29
+ * trick questions
30
+
31
+ # Results
32
 
33
  ![image/png](https://huggingface.co/datasets/froggeric/creativity/resolve/main/benchmark_results_2024-02-26.png)
34
 
35
+ # Remarks about some of the models
36
+
37
+
38
+ # Questions type
39
+
40
+ I will not provide the exact questions are used for various reasons, but I can provide some general ideas about which areas they cover:
41
+
42
+
43
+ # Other interesting benchmarks
44
+
45
  - https://eqbench.com/
46
  - https://chat.lmsys.org/