Update evals/1-mmlu.out
Browse files- evals/1-mmlu.out +53 -55
evals/1-mmlu.out
CHANGED
@@ -1,75 +1,73 @@
|
|
1 |
-
|
2 |
-
hf (pretrained=../,trust_remote_code=True), gen_kwargs: (None), limit: None, num_fewshot: 5, batch_size: 4
|
3 |
| Tasks |Version|Filter|n-shot|Metric| |Value | |Stderr|
|
4 |
|---------------------------------------|------:|------|-----:|------|---|-----:|---|-----:|
|
5 |
-
|mmlu | 2|none | |acc |↑ |0.
|
6 |
-
| - humanities | 2|none | |acc |↑ |0.
|
7 |
-
| - formal_logic | 1|none | 5|acc |↑ |0.
|
8 |
-
| - high_school_european_history | 1|none | 5|acc |↑ |0.
|
9 |
-
| - high_school_us_history | 1|none | 5|acc |↑ |0.
|
10 |
-
| - high_school_world_history | 1|none | 5|acc |↑ |0.
|
11 |
| - international_law | 1|none | 5|acc |↑ |0.7769|± |0.0380|
|
12 |
| - jurisprudence | 1|none | 5|acc |↑ |0.7870|± |0.0396|
|
13 |
| - logical_fallacies | 1|none | 5|acc |↑ |0.7546|± |0.0338|
|
14 |
| - moral_disputes | 1|none | 5|acc |↑ |0.7457|± |0.0234|
|
15 |
-
| - moral_scenarios | 1|none | 5|acc |↑ |0.
|
16 |
-
| - philosophy | 1|none | 5|acc |↑ |0.
|
17 |
-
| - prehistory | 1|none | 5|acc |↑ |0.
|
18 |
-
| - professional_law | 1|none | 5|acc |↑ |0.
|
19 |
-
| - world_religions | 1|none | 5|acc |↑ |0.
|
20 |
-
| - other | 2|none | |acc |↑ |0.
|
21 |
-
| - business_ethics | 1|none | 5|acc |↑ |0.
|
22 |
-
| - clinical_knowledge | 1|none | 5|acc |↑ |0.
|
23 |
-
| - college_medicine | 1|none | 5|acc |↑ |0.
|
24 |
-
| - global_facts | 1|none | 5|acc |↑ |0.
|
25 |
-
| - human_aging | 1|none | 5|acc |↑ |0.
|
26 |
| - management | 1|none | 5|acc |↑ |0.8155|± |0.0384|
|
27 |
| - marketing | 1|none | 5|acc |↑ |0.8974|± |0.0199|
|
28 |
-
| - medical_genetics | 1|none | 5|acc |↑ |0.
|
29 |
-
| - miscellaneous | 1|none | 5|acc |↑ |0.
|
30 |
-
| - nutrition | 1|none | 5|acc |↑ |0.
|
31 |
-
| - professional_accounting | 1|none | 5|acc |↑ |0.
|
32 |
-
| - professional_medicine | 1|none | 5|acc |↑ |0.
|
33 |
| - virology | 1|none | 5|acc |↑ |0.5422|± |0.0388|
|
34 |
-
| - social sciences | 2|none | |acc |↑ |0.
|
35 |
-
| - econometrics | 1|none | 5|acc |↑ |0.
|
36 |
| - high_school_geography | 1|none | 5|acc |↑ |0.8283|± |0.0269|
|
37 |
| - high_school_government_and_politics| 1|none | 5|acc |↑ |0.9171|± |0.0199|
|
38 |
-
| - high_school_macroeconomics | 1|none | 5|acc |↑ |0.
|
39 |
-
| - high_school_microeconomics | 1|none | 5|acc |↑ |0.
|
40 |
| - high_school_psychology | 1|none | 5|acc |↑ |0.8587|± |0.0149|
|
41 |
-
| - human_sexuality | 1|none | 5|acc |↑ |0.
|
42 |
| - professional_psychology | 1|none | 5|acc |↑ |0.7026|± |0.0185|
|
43 |
-
| - public_relations | 1|none | 5|acc |↑ |0.
|
44 |
-
| - security_studies | 1|none | 5|acc |↑ |0.
|
45 |
-
| - sociology | 1|none | 5|acc |↑ |0.
|
46 |
| - us_foreign_policy | 1|none | 5|acc |↑ |0.8400|± |0.0368|
|
47 |
-
| - stem | 2|none | |acc |↑ |0.
|
48 |
| - abstract_algebra | 1|none | 5|acc |↑ |0.3700|± |0.0485|
|
49 |
| - anatomy | 1|none | 5|acc |↑ |0.7037|± |0.0394|
|
50 |
-
| - astronomy | 1|none | 5|acc |↑ |0.
|
51 |
-
| - college_biology | 1|none | 5|acc |↑ |0.
|
52 |
-
| - college_chemistry | 1|none | 5|acc |↑ |0.
|
53 |
-
| - college_computer_science | 1|none | 5|acc |↑ |0.
|
54 |
-
| - college_mathematics | 1|none | 5|acc |↑ |0.
|
55 |
-
| - college_physics | 1|none | 5|acc |↑ |0.
|
56 |
-
| - computer_security | 1|none | 5|acc |↑ |0.
|
57 |
| - conceptual_physics | 1|none | 5|acc |↑ |0.5957|± |0.0321|
|
58 |
-
| - electrical_engineering | 1|none | 5|acc |↑ |0.
|
59 |
-
| - elementary_mathematics | 1|none | 5|acc |↑ |0.
|
60 |
-
| - high_school_biology | 1|none | 5|acc |↑ |0.
|
61 |
-
| - high_school_chemistry | 1|none | 5|acc |↑ |0.
|
62 |
-
| - high_school_computer_science | 1|none | 5|acc |↑ |0.
|
63 |
-
| - high_school_mathematics | 1|none | 5|acc |↑ |0.
|
64 |
-
| - high_school_physics | 1|none | 5|acc |↑ |0.
|
65 |
-
| - high_school_statistics | 1|none | 5|acc |↑ |0.
|
66 |
-
| - machine_learning | 1|none | 5|acc |↑ |0.
|
67 |
|
68 |
| Groups |Version|Filter|n-shot|Metric| |Value | |Stderr|
|
69 |
|------------------|------:|------|------|------|---|-----:|---|-----:|
|
70 |
-
|mmlu | 2|none | |acc |↑ |0.
|
71 |
-
| - humanities | 2|none | |acc |↑ |0.
|
72 |
-
| - other | 2|none | |acc |↑ |0.
|
73 |
-
| - social sciences| 2|none | |acc |↑ |0.
|
74 |
-
| - stem | 2|none | |acc |↑ |0.
|
75 |
-
|
|
|
1 |
+
f (pretrained=../,trust_remote_code=True), gen_kwargs: (None), limit: None, num_fewshot: 5, batch_size: 16
|
|
|
2 |
| Tasks |Version|Filter|n-shot|Metric| |Value | |Stderr|
|
3 |
|---------------------------------------|------:|------|-----:|------|---|-----:|---|-----:|
|
4 |
+
|mmlu | 2|none | |acc |↑ |0.6634|± |0.0038|
|
5 |
+
| - humanities | 2|none | |acc |↑ |0.6077|± |0.0068|
|
6 |
+
| - formal_logic | 1|none | 5|acc |↑ |0.5238|± |0.0447|
|
7 |
+
| - high_school_european_history | 1|none | 5|acc |↑ |0.7455|± |0.0340|
|
8 |
+
| - high_school_us_history | 1|none | 5|acc |↑ |0.8480|± |0.0252|
|
9 |
+
| - high_school_world_history | 1|none | 5|acc |↑ |0.8312|± |0.0244|
|
10 |
| - international_law | 1|none | 5|acc |↑ |0.7769|± |0.0380|
|
11 |
| - jurisprudence | 1|none | 5|acc |↑ |0.7870|± |0.0396|
|
12 |
| - logical_fallacies | 1|none | 5|acc |↑ |0.7546|± |0.0338|
|
13 |
| - moral_disputes | 1|none | 5|acc |↑ |0.7457|± |0.0234|
|
14 |
+
| - moral_scenarios | 1|none | 5|acc |↑ |0.4458|± |0.0166|
|
15 |
+
| - philosophy | 1|none | 5|acc |↑ |0.7460|± |0.0247|
|
16 |
+
| - prehistory | 1|none | 5|acc |↑ |0.7407|± |0.0244|
|
17 |
+
| - professional_law | 1|none | 5|acc |↑ |0.4752|± |0.0128|
|
18 |
+
| - world_religions | 1|none | 5|acc |↑ |0.8187|± |0.0295|
|
19 |
+
| - other | 2|none | |acc |↑ |0.7441|± |0.0076|
|
20 |
+
| - business_ethics | 1|none | 5|acc |↑ |0.6800|± |0.0469|
|
21 |
+
| - clinical_knowledge | 1|none | 5|acc |↑ |0.7509|± |0.0266|
|
22 |
+
| - college_medicine | 1|none | 5|acc |↑ |0.6647|± |0.0360|
|
23 |
+
| - global_facts | 1|none | 5|acc |↑ |0.4700|± |0.0502|
|
24 |
+
| - human_aging | 1|none | 5|acc |↑ |0.7265|± |0.0299|
|
25 |
| - management | 1|none | 5|acc |↑ |0.8155|± |0.0384|
|
26 |
| - marketing | 1|none | 5|acc |↑ |0.8974|± |0.0199|
|
27 |
+
| - medical_genetics | 1|none | 5|acc |↑ |0.8300|± |0.0378|
|
28 |
+
| - miscellaneous | 1|none | 5|acc |↑ |0.8519|± |0.0127|
|
29 |
+
| - nutrition | 1|none | 5|acc |↑ |0.7582|± |0.0245|
|
30 |
+
| - professional_accounting | 1|none | 5|acc |↑ |0.5426|± |0.0297|
|
31 |
+
| - professional_medicine | 1|none | 5|acc |↑ |0.7426|± |0.0266|
|
32 |
| - virology | 1|none | 5|acc |↑ |0.5422|± |0.0388|
|
33 |
+
| - social sciences | 2|none | |acc |↑ |0.7595|± |0.0075|
|
34 |
+
| - econometrics | 1|none | 5|acc |↑ |0.5263|± |0.0470|
|
35 |
| - high_school_geography | 1|none | 5|acc |↑ |0.8283|± |0.0269|
|
36 |
| - high_school_government_and_politics| 1|none | 5|acc |↑ |0.9171|± |0.0199|
|
37 |
+
| - high_school_macroeconomics | 1|none | 5|acc |↑ |0.6615|± |0.0240|
|
38 |
+
| - high_school_microeconomics | 1|none | 5|acc |↑ |0.7395|± |0.0285|
|
39 |
| - high_school_psychology | 1|none | 5|acc |↑ |0.8587|± |0.0149|
|
40 |
+
| - human_sexuality | 1|none | 5|acc |↑ |0.7786|± |0.0364|
|
41 |
| - professional_psychology | 1|none | 5|acc |↑ |0.7026|± |0.0185|
|
42 |
+
| - public_relations | 1|none | 5|acc |↑ |0.6545|± |0.0455|
|
43 |
+
| - security_studies | 1|none | 5|acc |↑ |0.7143|± |0.0289|
|
44 |
+
| - sociology | 1|none | 5|acc |↑ |0.8507|± |0.0252|
|
45 |
| - us_foreign_policy | 1|none | 5|acc |↑ |0.8400|± |0.0368|
|
46 |
+
| - stem | 2|none | |acc |↑ |0.5734|± |0.0084|
|
47 |
| - abstract_algebra | 1|none | 5|acc |↑ |0.3700|± |0.0485|
|
48 |
| - anatomy | 1|none | 5|acc |↑ |0.7037|± |0.0394|
|
49 |
+
| - astronomy | 1|none | 5|acc |↑ |0.7566|± |0.0349|
|
50 |
+
| - college_biology | 1|none | 5|acc |↑ |0.8264|± |0.0317|
|
51 |
+
| - college_chemistry | 1|none | 5|acc |↑ |0.4800|± |0.0502|
|
52 |
+
| - college_computer_science | 1|none | 5|acc |↑ |0.5300|± |0.0502|
|
53 |
+
| - college_mathematics | 1|none | 5|acc |↑ |0.3500|± |0.0479|
|
54 |
+
| - college_physics | 1|none | 5|acc |↑ |0.4902|± |0.0497|
|
55 |
+
| - computer_security | 1|none | 5|acc |↑ |0.7600|± |0.0429|
|
56 |
| - conceptual_physics | 1|none | 5|acc |↑ |0.5957|± |0.0321|
|
57 |
+
| - electrical_engineering | 1|none | 5|acc |↑ |0.6069|± |0.0407|
|
58 |
+
| - elementary_mathematics | 1|none | 5|acc |↑ |0.4656|± |0.0257|
|
59 |
+
| - high_school_biology | 1|none | 5|acc |↑ |0.8032|± |0.0226|
|
60 |
+
| - high_school_chemistry | 1|none | 5|acc |↑ |0.5567|± |0.0350|
|
61 |
+
| - high_school_computer_science | 1|none | 5|acc |↑ |0.6900|± |0.0465|
|
62 |
+
| - high_school_mathematics | 1|none | 5|acc |↑ |0.3704|± |0.0294|
|
63 |
+
| - high_school_physics | 1|none | 5|acc |↑ |0.4702|± |0.0408|
|
64 |
+
| - high_school_statistics | 1|none | 5|acc |↑ |0.5694|± |0.0338|
|
65 |
+
| - machine_learning | 1|none | 5|acc |↑ |0.4554|± |0.0473|
|
66 |
|
67 |
| Groups |Version|Filter|n-shot|Metric| |Value | |Stderr|
|
68 |
|------------------|------:|------|------|------|---|-----:|---|-----:|
|
69 |
+
|mmlu | 2|none | |acc |↑ |0.6634|± |0.0038|
|
70 |
+
| - humanities | 2|none | |acc |↑ |0.6077|± |0.0068|
|
71 |
+
| - other | 2|none | |acc |↑ |0.7441|± |0.0076|
|
72 |
+
| - social sciences| 2|none | |acc |↑ |0.7595|± |0.0075|
|
73 |
+
| - stem | 2|none | |acc |↑ |0.5734|± |0.0084|
|
|