Update README.md
Browse files
README.md
CHANGED
@@ -33,18 +33,19 @@ model-index:
|
|
33 |
|
34 |
## Performance and Evaluation
|
35 |
|
36 |
-
|
37 |
-
|----------------|-----------|------------|-----------------------------------------------------|
|
38 |
-
| HellaSwag | acc | **0.291** | 0.289 |0.2829 |
|
39 |
-
| SciQ | acc | **0.754** | 0.752 |0.726|
|
40 |
-
| Winogrande | acc | 0.491 | **0.516** | 0.4909|
|
41 |
-
| TruthfulQA MC1 | acc | 0.236 | 0.228 | **0.2619** |
|
42 |
-
| MMLU (overall) | acc | 0.230 | 0.229 | **0.2310** |
|
43 |
-
| - Humanities | acc | 0.242 | 0.242 | **0.2387** |
|
44 |
-
| - Social Sci. | acc | 0.217 | 0.217 | **0.2246** |
|
45 |
-
| - STEM | acc | 0.213 | 0.213 | **0.2226** |
|
46 |
-
| - Other | acc | **0.239** | 0.238 | **0.2343** |
|
47 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
48 |
|
49 |
|
50 |
|
|
|
33 |
|
34 |
## Performance and Evaluation
|
35 |
|
36 |
+
## Performance and Evaluation
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
37 |
|
38 |
+
| Dataset | Metric | thecr7guy/gpt2-pretrain | GPT-2 (baseline) | thecr7guy/gpt2-insFT |
|
39 |
+
| ------------------ | ------ | ----------------------- | ---------------- | -------------------- |
|
40 |
+
| **HellaSwag** | acc | **0.291** | 0.289 | 0.2829 |
|
41 |
+
| **SciQ** | acc | **0.754** | 0.752 | 0.726 |
|
42 |
+
| **Winogrande** | acc | 0.491 | **0.516** | 0.4909 |
|
43 |
+
| **TruthfulQA MC1** | acc | 0.236 | 0.228 | **0.2619** |
|
44 |
+
| **MMLU (overall)** | acc | 0.230 | 0.229 | **0.2310** |
|
45 |
+
| ββ Humanities | acc | 0.242 | 0.242 | **0.2387** |
|
46 |
+
| ββ Social Sci. | acc | 0.217 | 0.217 | **0.2246** |
|
47 |
+
| ββ STEM | acc | 0.213 | 0.213 | **0.2226** |
|
48 |
+
| ββ Other | acc | 0.239 | 0.238 | **0.2343** |
|
49 |
|
50 |
|
51 |
|