Update README.md
Browse files
README.md
CHANGED
|
@@ -30,6 +30,48 @@ tags:
|
|
| 30 |
</p>
|
| 31 |
|
| 32 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
Prompt Template(Alpaca)
|
| 34 |
### Instruction:
|
| 35 |
{instruction}
|
|
|
|
| 30 |
</p>
|
| 31 |
|
| 32 |
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
########## First turn ##########
|
| 36 |
+
| model | turn | score |
|
| 37 |
+
|--------------------|------|----------|
|
| 38 |
+
| gpt-4 | 1 | 8.95625 |
|
| 39 |
+
| xDAN-L1-Chat-RL-v1 | 1 | 8.87500 |
|
| 40 |
+
| claude-v1 | 1 | 8.15000 |
|
| 41 |
+
| gpt-3.5-turbo | 1 | 8.07500 |
|
| 42 |
+
| claude-instant-v1 | 1 | 7.80000 |
|
| 43 |
+
| vicuna-33b-v1.3 | 1 | 7.45625 |
|
| 44 |
+
| wizardlm-30b | 1 | 7.13125 |
|
| 45 |
+
| oasst-sft-7-llama-30b | 1 | 7.10625 |
|
| 46 |
+
| Llama-2-70b-chat | 1 | 6.98750 |
|
| 47 |
+
|
| 48 |
+
########## Second turn ##########
|
| 49 |
+
| model | turn | score |
|
| 50 |
+
|--------------------|------|-----------|
|
| 51 |
+
| gpt-4 | 2 | 9.025000 |
|
| 52 |
+
| claude-instant-v1 | 2 | 8.012658 |
|
| 53 |
+
| xDAN-L1-Chat-RL-v1 | 2 | 7.825000 |
|
| 54 |
+
| gpt-3.5-turbo | 2 | 7.812500 |
|
| 55 |
+
| claude-v1 | 2 | 7.650000 |
|
| 56 |
+
| wizardlm-30b | 2 | 6.887500 |
|
| 57 |
+
| vicuna-33b-v1.3 | 2 | 6.787500 |
|
| 58 |
+
| Llama-2-70b-chat | 2 | 6.725000 |
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
########## Average turn##########
|
| 62 |
+
| model | score |
|
| 63 |
+
|--------------------|-----------|
|
| 64 |
+
| gpt-4 | 8.990625 |
|
| 65 |
+
| xDAN-L1-Chat-RL-v1 | 8.350000 |
|
| 66 |
+
| gpt-3.5-turbo | 7.943750 |
|
| 67 |
+
| claude-instant-v1 | 7.905660 |
|
| 68 |
+
| claude-v1 | 7.900000 |
|
| 69 |
+
| vicuna-33b-v1.3 | 7.121875 |
|
| 70 |
+
| wizardlm-30b | 7.009375 |
|
| 71 |
+
| Llama-2-70b-chat | 6.856250 |
|
| 72 |
+
|
| 73 |
+
|
| 74 |
+
|
| 75 |
Prompt Template(Alpaca)
|
| 76 |
### Instruction:
|
| 77 |
{instruction}
|