Update README.md
Browse files
README.md
CHANGED
@@ -28,8 +28,8 @@ Nous Hermes 2 Yi 34B was trained on 1,000,000 entries of primarily GPT-4 generat
|
|
28 |
|
29 |
# Table of Contents
|
30 |
1. [Example Outputs](#example-outputs)
|
|
|
31 |
- Create a Flask based FTP Server
|
32 |
-
-
|
33 |
3. [Benchmark Results](#benchmark-results)
|
34 |
- GPT4All
|
35 |
- AGIEval
|
@@ -55,6 +55,22 @@ Nous Hermes 2 Yi 34B was trained on 1,000,000 entries of primarily GPT-4 generat
|
|
55 |
|
56 |
Nous-Hermes 2 on Yi 34B outperforms all Nous-Hermes & Open-Hermes models of the past, achieving new heights in all benchmarks for a Nous Research LLM.
|
57 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
58 |
## GPT4All
|
59 |
GPT-4All Benchmark Set
|
60 |
```
|
|
|
28 |
|
29 |
# Table of Contents
|
30 |
1. [Example Outputs](#example-outputs)
|
31 |
+
- Discussing the Laws of Gravity
|
32 |
- Create a Flask based FTP Server
|
|
|
33 |
3. [Benchmark Results](#benchmark-results)
|
34 |
- GPT4All
|
35 |
- AGIEval
|
|
|
55 |
|
56 |
Nous-Hermes 2 on Yi 34B outperforms all Nous-Hermes & Open-Hermes models of the past, achieving new heights in all benchmarks for a Nous Research LLM.
|
57 |
|
58 |
+
# Benchmarks Compared
|
59 |
+
|
60 |
+
### GPT4All:
|
61 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6317aade83d8d2fd903192d9/91onORUcUrAqTb3b9mG5e.png)
|
62 |
+
|
63 |
+
### AGIEval:
|
64 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6317aade83d8d2fd903192d9/hqDpMlKpINfDf4PmB31uW.png)
|
65 |
+
|
66 |
+
### BigBench:
|
67 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6317aade83d8d2fd903192d9/uh8mZZg_wZinFysxcfLSF.png)
|
68 |
+
|
69 |
+
### TruthfulQA:
|
70 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6317aade83d8d2fd903192d9/N_cX6YAWjJsvClotuoPdH.png)
|
71 |
+
|
72 |
+
|
73 |
+
|
74 |
## GPT4All
|
75 |
GPT-4All Benchmark Set
|
76 |
```
|