Update README.md
Browse files
README.md
CHANGED
@@ -64,7 +64,6 @@ In AlpacaEval, Rocket 🦝 achieves a near 80% win rate, coupled with an average
|
|
64 |
|
65 |
|
66 |
## Other benchmarks
|
67 |
-
Despite its impressive performance on MT-Bench and AlpacaEval benchmarks, the model experiences some challenges when evaluated on other benchmark tests.
|
68 |
|
69 |
| Metric | Value |
|
70 |
|-----------------------|---------------------------|
|
@@ -72,9 +71,9 @@ Despite its impressive performance on MT-Bench and AlpacaEval benchmarks, the mo
|
|
72 |
| ARC (25-shot) | 50.51 |
|
73 |
| HellaSwag (10-shot) | 73.91 |
|
74 |
| MMLU (5-shot) | 61.07 |
|
75 |
-
| TruthfulQA (0-shot) |
|
76 |
| Winogrande (5-shot) | 63.22 |
|
77 |
-
| GSM8K (5-shot) |
|
78 |
| DROP (3-shot) | 9.66 |
|
79 |
|
80 |
|
|
|
64 |
|
65 |
|
66 |
## Other benchmarks
|
|
|
67 |
|
68 |
| Metric | Value |
|
69 |
|-----------------------|---------------------------|
|
|
|
71 |
| ARC (25-shot) | 50.51 |
|
72 |
| HellaSwag (10-shot) | 73.91 |
|
73 |
| MMLU (5-shot) | 61.07 |
|
74 |
+
| TruthfulQA (mc2) (0-shot) | 54.38 |
|
75 |
| Winogrande (5-shot) | 63.22 |
|
76 |
+
| GSM8K (5-shot) | 37.91 |
|
77 |
| DROP (3-shot) | 9.66 |
|
78 |
|
79 |
|