Update README.md
Browse files
README.md
CHANGED
|
@@ -18,7 +18,18 @@ The feedback encompasses a wide range of input documents, from short to lengthy
|
|
| 18 |
- Four non-dialouge domains: News, Lifestyle, Report, Medical
|
| 19 |
- Three dialogue domains: Daily Life, Interview, Meeting
|
| 20 |
|
| 21 |
-
Surprisingly, it outperforms the nearly 10x larger Llama3-70B-Instruct while offering much faster inference speed.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
|
| 23 |
This is human evaluation results:
|
| 24 |
|
|
|
|
| 18 |
- Four non-dialouge domains: News, Lifestyle, Report, Medical
|
| 19 |
- Three dialogue domains: Daily Life, Interview, Meeting
|
| 20 |
|
| 21 |
+
Surprisingly, it outperforms the nearly 10x larger **Llama3-70B-Instruct** and also **GPT-4o** while offering much faster inference speed.
|
| 22 |
+
|
| 23 |
+
This is automated evaluation results:
|
| 24 |
+
|
| 25 |
+
| **Config.** | **Faithfulness** | **Completeness** | **Conciseness** | **Average** |
|
| 26 |
+
|--------------------|------------|-----------|-----------|----------|
|
| 27 |
+
| Llama3-8B-Instruct | 0.864 | 0.583 | 0.450 | 0.632 |
|
| 28 |
+
| Llama3-70B-Instruct | 0.931 | 0.596 | 0.487 | 0.671 |
|
| 29 |
+
| GPT-4o | 0.940 | 0.657 | 0.437 | 0.678 |
|
| 30 |
+
| SummLlama3-8B | 0.931 | 0.614 | 0.659 | 0.735 |
|
| 31 |
+
| SummLlama3-70B | 0.950 | 0.632 | 0.754 | 0.779 |
|
| 32 |
+
|
| 33 |
|
| 34 |
This is human evaluation results:
|
| 35 |
|