Hugging Face's logo

Datasets:
michaelchenkj
/
test-models

Dataset card Files
xet
 Community
michaelchenkj commited on
Commit
2fc8c43
·
verified ·
1 Parent(s): ce3e7be

Upload folder using huggingface_hub

Browse files
Files changed (11) hide show
  1. results/baseline-lr6e-4/checkpoint-1080.json +96 -0
  2. results/baseline-lr6e-4/checkpoint-1440.json +96 -0
  3. results/baseline-lr6e-4/checkpoint-1800.json +96 -0
  4. results/baseline-lr6e-4/checkpoint-2160.json +96 -0
  5. results/baseline-lr6e-4/checkpoint-2520.json +96 -0
  6. results/baseline-lr6e-4/checkpoint-2880.json +96 -0
  7. results/baseline-lr6e-4/checkpoint-3240.json +96 -0
  8. results/baseline-lr6e-4/checkpoint-360.json +100 -0
  9. results/baseline-lr6e-4/checkpoint-3600.json +96 -0
  10. results/baseline-lr6e-4/checkpoint-72.json +118 -0
  11. results/baseline-lr6e-4/checkpoint-720.json +96 -0
results/baseline-lr6e-4/checkpoint-1080.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_path": "/workspace/concept_model/runs/baseline-lr6e-4/checkpoint-1080",
3
+ "run": "baseline-lr6e-4",
4
+ "checkpoint": "checkpoint-1080",
5
+ "tasks": [
6
+ {
7
+ "task": "arc-challenge",
8
+ "accuracies": [
9
+ 0.2652360515021459
10
+ ],
11
+ "mean": 0.2652360515021459,
12
+ "ci": 0.0
13
+ },
14
+ {
15
+ "task": "arc-easy",
16
+ "accuracies": [
17
+ 0.26765327695560254
18
+ ],
19
+ "mean": 0.26765327695560254,
20
+ "ci": 0.0
21
+ },
22
+ {
23
+ "task": "boolq",
24
+ "accuracies": [
25
+ 0.6220183486238532
26
+ ],
27
+ "mean": 0.6220183486238532,
28
+ "ci": 0.0
29
+ },
30
+ {
31
+ "task": "commonsenseqa",
32
+ "accuracies": [
33
+ 0.19737919737919737
34
+ ],
35
+ "mean": 0.19737919737919737,
36
+ "ci": 0.0
37
+ },
38
+ {
39
+ "task": "hellaswag",
40
+ "accuracies": [
41
+ 0.25731925911173076
42
+ ],
43
+ "mean": 0.25731925911173076,
44
+ "ci": 0.0
45
+ },
46
+ {
47
+ "task": "logiqa",
48
+ "accuracies": [
49
+ 0.2749615975422427
50
+ ],
51
+ "mean": 0.2749615975422427,
52
+ "ci": 0.0
53
+ },
54
+ {
55
+ "task": "mathqa",
56
+ "accuracies": [
57
+ 0.22646566164154103
58
+ ],
59
+ "mean": 0.22646566164154103,
60
+ "ci": 0.0
61
+ },
62
+ {
63
+ "task": "openbookqa",
64
+ "accuracies": [
65
+ 0.264
66
+ ],
67
+ "mean": 0.264,
68
+ "ci": 0.0
69
+ },
70
+ {
71
+ "task": "piqa",
72
+ "accuracies": [
73
+ 0.49891186071817195
74
+ ],
75
+ "mean": 0.49891186071817195,
76
+ "ci": 0.0
77
+ },
78
+ {
79
+ "task": "pubmedqa",
80
+ "accuracies": [
81
+ 0.545
82
+ ],
83
+ "mean": 0.545,
84
+ "ci": 0.0
85
+ },
86
+ {
87
+ "task": "winogrande",
88
+ "accuracies": [
89
+ 0.5114443567482242
90
+ ],
91
+ "mean": 0.5114443567482242,
92
+ "ci": 0.0
93
+ }
94
+ ],
95
+ "mean_accuracy": 0.35730814638388275
96
+ }
results/baseline-lr6e-4/checkpoint-1440.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_path": "/workspace/concept_model/runs/baseline-lr6e-4/checkpoint-1440",
3
+ "run": "baseline-lr6e-4",
4
+ "checkpoint": "checkpoint-1440",
5
+ "tasks": [
6
+ {
7
+ "task": "arc-challenge",
8
+ "accuracies": [
9
+ 0.2652360515021459
10
+ ],
11
+ "mean": 0.2652360515021459,
12
+ "ci": 0.0
13
+ },
14
+ {
15
+ "task": "arc-easy",
16
+ "accuracies": [
17
+ 0.26807610993657505
18
+ ],
19
+ "mean": 0.26807610993657505,
20
+ "ci": 0.0
21
+ },
22
+ {
23
+ "task": "boolq",
24
+ "accuracies": [
25
+ 0.6229357798165137
26
+ ],
27
+ "mean": 0.6229357798165137,
28
+ "ci": 0.0
29
+ },
30
+ {
31
+ "task": "commonsenseqa",
32
+ "accuracies": [
33
+ 0.2022932022932023
34
+ ],
35
+ "mean": 0.2022932022932023,
36
+ "ci": 0.0
37
+ },
38
+ {
39
+ "task": "hellaswag",
40
+ "accuracies": [
41
+ 0.2608046205935073
42
+ ],
43
+ "mean": 0.2608046205935073,
44
+ "ci": 0.0
45
+ },
46
+ {
47
+ "task": "logiqa",
48
+ "accuracies": [
49
+ 0.2749615975422427
50
+ ],
51
+ "mean": 0.2749615975422427,
52
+ "ci": 0.0
53
+ },
54
+ {
55
+ "task": "mathqa",
56
+ "accuracies": [
57
+ 0.22680067001675042
58
+ ],
59
+ "mean": 0.22680067001675042,
60
+ "ci": 0.0
61
+ },
62
+ {
63
+ "task": "openbookqa",
64
+ "accuracies": [
65
+ 0.262
66
+ ],
67
+ "mean": 0.262,
68
+ "ci": 0.0
69
+ },
70
+ {
71
+ "task": "piqa",
72
+ "accuracies": [
73
+ 0.515233949945593
74
+ ],
75
+ "mean": 0.515233949945593,
76
+ "ci": 0.0
77
+ },
78
+ {
79
+ "task": "pubmedqa",
80
+ "accuracies": [
81
+ 0.553
82
+ ],
83
+ "mean": 0.553,
84
+ "ci": 0.0
85
+ },
86
+ {
87
+ "task": "winogrande",
88
+ "accuracies": [
89
+ 0.48855564325177586
90
+ ],
91
+ "mean": 0.48855564325177586,
92
+ "ci": 0.0
93
+ }
94
+ ],
95
+ "mean_accuracy": 0.3581725113543915
96
+ }
results/baseline-lr6e-4/checkpoint-1800.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_path": "/workspace/concept_model/runs/baseline-lr6e-4/checkpoint-1800",
3
+ "run": "baseline-lr6e-4",
4
+ "checkpoint": "checkpoint-1800",
5
+ "tasks": [
6
+ {
7
+ "task": "arc-challenge",
8
+ "accuracies": [
9
+ 0.2652360515021459
10
+ ],
11
+ "mean": 0.2652360515021459,
12
+ "ci": 0.0
13
+ },
14
+ {
15
+ "task": "arc-easy",
16
+ "accuracies": [
17
+ 0.26765327695560254
18
+ ],
19
+ "mean": 0.26765327695560254,
20
+ "ci": 0.0
21
+ },
22
+ {
23
+ "task": "boolq",
24
+ "accuracies": [
25
+ 0.6226299694189602
26
+ ],
27
+ "mean": 0.6226299694189602,
28
+ "ci": 0.0
29
+ },
30
+ {
31
+ "task": "commonsenseqa",
32
+ "accuracies": [
33
+ 0.2072072072072072
34
+ ],
35
+ "mean": 0.2072072072072072,
36
+ "ci": 0.0
37
+ },
38
+ {
39
+ "task": "hellaswag",
40
+ "accuracies": [
41
+ 0.2577175861382195
42
+ ],
43
+ "mean": 0.2577175861382195,
44
+ "ci": 0.0
45
+ },
46
+ {
47
+ "task": "logiqa",
48
+ "accuracies": [
49
+ 0.2749615975422427
50
+ ],
51
+ "mean": 0.2749615975422427,
52
+ "ci": 0.0
53
+ },
54
+ {
55
+ "task": "mathqa",
56
+ "accuracies": [
57
+ 0.22680067001675042
58
+ ],
59
+ "mean": 0.22680067001675042,
60
+ "ci": 0.0
61
+ },
62
+ {
63
+ "task": "openbookqa",
64
+ "accuracies": [
65
+ 0.268
66
+ ],
67
+ "mean": 0.268,
68
+ "ci": 0.0
69
+ },
70
+ {
71
+ "task": "piqa",
72
+ "accuracies": [
73
+ 0.5054406964091404
74
+ ],
75
+ "mean": 0.5054406964091404,
76
+ "ci": 0.0
77
+ },
78
+ {
79
+ "task": "pubmedqa",
80
+ "accuracies": [
81
+ 0.547
82
+ ],
83
+ "mean": 0.547,
84
+ "ci": 0.0
85
+ },
86
+ {
87
+ "task": "winogrande",
88
+ "accuracies": [
89
+ 0.5019731649565904
90
+ ],
91
+ "mean": 0.5019731649565904,
92
+ "ci": 0.0
93
+ }
94
+ ],
95
+ "mean_accuracy": 0.35860183819516905
96
+ }
results/baseline-lr6e-4/checkpoint-2160.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_path": "/workspace/concept_model/runs/baseline-lr6e-4/checkpoint-2160",
3
+ "run": "baseline-lr6e-4",
4
+ "checkpoint": "checkpoint-2160",
5
+ "tasks": [
6
+ {
7
+ "task": "arc-challenge",
8
+ "accuracies": [
9
+ 0.2652360515021459
10
+ ],
11
+ "mean": 0.2652360515021459,
12
+ "ci": 0.0
13
+ },
14
+ {
15
+ "task": "arc-easy",
16
+ "accuracies": [
17
+ 0.2693446088794926
18
+ ],
19
+ "mean": 0.2693446088794926,
20
+ "ci": 0.0
21
+ },
22
+ {
23
+ "task": "boolq",
24
+ "accuracies": [
25
+ 0.618348623853211
26
+ ],
27
+ "mean": 0.618348623853211,
28
+ "ci": 0.0
29
+ },
30
+ {
31
+ "task": "commonsenseqa",
32
+ "accuracies": [
33
+ 0.20065520065520065
34
+ ],
35
+ "mean": 0.20065520065520065,
36
+ "ci": 0.0
37
+ },
38
+ {
39
+ "task": "hellaswag",
40
+ "accuracies": [
41
+ 0.2582154949213304
42
+ ],
43
+ "mean": 0.2582154949213304,
44
+ "ci": 0.0
45
+ },
46
+ {
47
+ "task": "logiqa",
48
+ "accuracies": [
49
+ 0.2749615975422427
50
+ ],
51
+ "mean": 0.2749615975422427,
52
+ "ci": 0.0
53
+ },
54
+ {
55
+ "task": "mathqa",
56
+ "accuracies": [
57
+ 0.2234505862646566
58
+ ],
59
+ "mean": 0.2234505862646566,
60
+ "ci": 0.0
61
+ },
62
+ {
63
+ "task": "openbookqa",
64
+ "accuracies": [
65
+ 0.278
66
+ ],
67
+ "mean": 0.278,
68
+ "ci": 0.0
69
+ },
70
+ {
71
+ "task": "piqa",
72
+ "accuracies": [
73
+ 0.5038084874863983
74
+ ],
75
+ "mean": 0.5038084874863983,
76
+ "ci": 0.0
77
+ },
78
+ {
79
+ "task": "pubmedqa",
80
+ "accuracies": [
81
+ 0.539
82
+ ],
83
+ "mean": 0.539,
84
+ "ci": 0.0
85
+ },
86
+ {
87
+ "task": "winogrande",
88
+ "accuracies": [
89
+ 0.48697711128650356
90
+ ],
91
+ "mean": 0.48697711128650356,
92
+ "ci": 0.0
93
+ }
94
+ ],
95
+ "mean_accuracy": 0.3561816147628347
96
+ }
results/baseline-lr6e-4/checkpoint-2520.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_path": "/workspace/concept_model/runs/baseline-lr6e-4/checkpoint-2520",
3
+ "run": "baseline-lr6e-4",
4
+ "checkpoint": "checkpoint-2520",
5
+ "tasks": [
6
+ {
7
+ "task": "arc-challenge",
8
+ "accuracies": [
9
+ 0.2652360515021459
10
+ ],
11
+ "mean": 0.2652360515021459,
12
+ "ci": 0.0
13
+ },
14
+ {
15
+ "task": "arc-easy",
16
+ "accuracies": [
17
+ 0.2714587737843552
18
+ ],
19
+ "mean": 0.2714587737843552,
20
+ "ci": 0.0
21
+ },
22
+ {
23
+ "task": "boolq",
24
+ "accuracies": [
25
+ 0.618960244648318
26
+ ],
27
+ "mean": 0.618960244648318,
28
+ "ci": 0.0
29
+ },
30
+ {
31
+ "task": "commonsenseqa",
32
+ "accuracies": [
33
+ 0.2113022113022113
34
+ ],
35
+ "mean": 0.2113022113022113,
36
+ "ci": 0.0
37
+ },
38
+ {
39
+ "task": "hellaswag",
40
+ "accuracies": [
41
+ 0.2590121489743079
42
+ ],
43
+ "mean": 0.2590121489743079,
44
+ "ci": 0.0
45
+ },
46
+ {
47
+ "task": "logiqa",
48
+ "accuracies": [
49
+ 0.28110599078341014
50
+ ],
51
+ "mean": 0.28110599078341014,
52
+ "ci": 0.0
53
+ },
54
+ {
55
+ "task": "mathqa",
56
+ "accuracies": [
57
+ 0.22613065326633167
58
+ ],
59
+ "mean": 0.22613065326633167,
60
+ "ci": 0.0
61
+ },
62
+ {
63
+ "task": "openbookqa",
64
+ "accuracies": [
65
+ 0.276
66
+ ],
67
+ "mean": 0.276,
68
+ "ci": 0.0
69
+ },
70
+ {
71
+ "task": "piqa",
72
+ "accuracies": [
73
+ 0.5059847660500544
74
+ ],
75
+ "mean": 0.5059847660500544,
76
+ "ci": 0.0
77
+ },
78
+ {
79
+ "task": "pubmedqa",
80
+ "accuracies": [
81
+ 0.512
82
+ ],
83
+ "mean": 0.512,
84
+ "ci": 0.0
85
+ },
86
+ {
87
+ "task": "winogrande",
88
+ "accuracies": [
89
+ 0.49171270718232046
90
+ ],
91
+ "mean": 0.49171270718232046,
92
+ "ci": 0.0
93
+ }
94
+ ],
95
+ "mean_accuracy": 0.35626395886304135
96
+ }
results/baseline-lr6e-4/checkpoint-2880.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_path": "/workspace/concept_model/runs/baseline-lr6e-4/checkpoint-2880",
3
+ "run": "baseline-lr6e-4",
4
+ "checkpoint": "checkpoint-2880",
5
+ "tasks": [
6
+ {
7
+ "task": "arc-challenge",
8
+ "accuracies": [
9
+ 0.26437768240343346
10
+ ],
11
+ "mean": 0.26437768240343346,
12
+ "ci": 0.0
13
+ },
14
+ {
15
+ "task": "arc-easy",
16
+ "accuracies": [
17
+ 0.26976744186046514
18
+ ],
19
+ "mean": 0.26976744186046514,
20
+ "ci": 0.0
21
+ },
22
+ {
23
+ "task": "boolq",
24
+ "accuracies": [
25
+ 0.6113149847094801
26
+ ],
27
+ "mean": 0.6113149847094801,
28
+ "ci": 0.0
29
+ },
30
+ {
31
+ "task": "commonsenseqa",
32
+ "accuracies": [
33
+ 0.20638820638820637
34
+ ],
35
+ "mean": 0.20638820638820637,
36
+ "ci": 0.0
37
+ },
38
+ {
39
+ "task": "hellaswag",
40
+ "accuracies": [
41
+ 0.25951005775741887
42
+ ],
43
+ "mean": 0.25951005775741887,
44
+ "ci": 0.0
45
+ },
46
+ {
47
+ "task": "logiqa",
48
+ "accuracies": [
49
+ 0.282642089093702
50
+ ],
51
+ "mean": 0.282642089093702,
52
+ "ci": 0.0
53
+ },
54
+ {
55
+ "task": "mathqa",
56
+ "accuracies": [
57
+ 0.2221105527638191
58
+ ],
59
+ "mean": 0.2221105527638191,
60
+ "ci": 0.0
61
+ },
62
+ {
63
+ "task": "openbookqa",
64
+ "accuracies": [
65
+ 0.274
66
+ ],
67
+ "mean": 0.274,
68
+ "ci": 0.0
69
+ },
70
+ {
71
+ "task": "piqa",
72
+ "accuracies": [
73
+ 0.49020674646354734
74
+ ],
75
+ "mean": 0.49020674646354734,
76
+ "ci": 0.0
77
+ },
78
+ {
79
+ "task": "pubmedqa",
80
+ "accuracies": [
81
+ 0.522
82
+ ],
83
+ "mean": 0.522,
84
+ "ci": 0.0
85
+ },
86
+ {
87
+ "task": "winogrande",
88
+ "accuracies": [
89
+ 0.49171270718232046
90
+ ],
91
+ "mean": 0.49171270718232046,
92
+ "ci": 0.0
93
+ }
94
+ ],
95
+ "mean_accuracy": 0.35400276987476303
96
+ }
results/baseline-lr6e-4/checkpoint-3240.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_path": "/workspace/concept_model/runs/baseline-lr6e-4/checkpoint-3240",
3
+ "run": "baseline-lr6e-4",
4
+ "checkpoint": "checkpoint-3240",
5
+ "tasks": [
6
+ {
7
+ "task": "arc-challenge",
8
+ "accuracies": [
9
+ 0.26609442060085836
10
+ ],
11
+ "mean": 0.26609442060085836,
12
+ "ci": 0.0
13
+ },
14
+ {
15
+ "task": "arc-easy",
16
+ "accuracies": [
17
+ 0.2617336152219873
18
+ ],
19
+ "mean": 0.2617336152219873,
20
+ "ci": 0.0
21
+ },
22
+ {
23
+ "task": "boolq",
24
+ "accuracies": [
25
+ 0.609480122324159
26
+ ],
27
+ "mean": 0.609480122324159,
28
+ "ci": 0.0
29
+ },
30
+ {
31
+ "task": "commonsenseqa",
32
+ "accuracies": [
33
+ 0.2113022113022113
34
+ ],
35
+ "mean": 0.2113022113022113,
36
+ "ci": 0.0
37
+ },
38
+ {
39
+ "task": "hellaswag",
40
+ "accuracies": [
41
+ 0.25941047600079664
42
+ ],
43
+ "mean": 0.25941047600079664,
44
+ "ci": 0.0
45
+ },
46
+ {
47
+ "task": "logiqa",
48
+ "accuracies": [
49
+ 0.271889400921659
50
+ ],
51
+ "mean": 0.271889400921659,
52
+ "ci": 0.0
53
+ },
54
+ {
55
+ "task": "mathqa",
56
+ "accuracies": [
57
+ 0.21139028475711893
58
+ ],
59
+ "mean": 0.21139028475711893,
60
+ "ci": 0.0
61
+ },
62
+ {
63
+ "task": "openbookqa",
64
+ "accuracies": [
65
+ 0.278
66
+ ],
67
+ "mean": 0.278,
68
+ "ci": 0.0
69
+ },
70
+ {
71
+ "task": "piqa",
72
+ "accuracies": [
73
+ 0.4967355821545158
74
+ ],
75
+ "mean": 0.4967355821545158,
76
+ "ci": 0.0
77
+ },
78
+ {
79
+ "task": "pubmedqa",
80
+ "accuracies": [
81
+ 0.521
82
+ ],
83
+ "mean": 0.521,
84
+ "ci": 0.0
85
+ },
86
+ {
87
+ "task": "winogrande",
88
+ "accuracies": [
89
+ 0.49013417521704816
90
+ ],
91
+ "mean": 0.49013417521704816,
92
+ "ci": 0.0
93
+ }
94
+ ],
95
+ "mean_accuracy": 0.352470026227305
96
+ }
results/baseline-lr6e-4/checkpoint-360.json ADDED
@@ -0,0 +1,100 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_path": "/workspace/concept_model/runs/baseline-lr6e-4/checkpoint-360",
3
+ "run": "baseline-lr6e-4",
4
+ "checkpoint": "checkpoint-360",
5
+ "tasks": [
6
+ {
7
+ "task": "arc-challenge",
8
+ "accuracies": [
9
+ 0.2652360515021459,
10
+ 0.2652360515021459,
11
+ 0.2652360515021459
12
+ ],
13
+ "mean": 0.2652360515021459,
14
+ "ci": NaN
15
+ },
16
+ {
17
+ "task": "arc-easy",
18
+ "accuracies": [
19
+ 0.26765327695560254,
20
+ 0.26765327695560254,
21
+ 0.26765327695560254
22
+ ],
23
+ "mean": 0.26765327695560254,
24
+ "ci": NaN
25
+ },
26
+ {
27
+ "task": "boolq",
28
+ "accuracies": [
29
+ 0.3782874617737003
30
+ ],
31
+ "mean": 0.3782874617737003,
32
+ "ci": 0.0
33
+ },
34
+ {
35
+ "task": "commonsenseqa",
36
+ "accuracies": [
37
+ 0.19737919737919737
38
+ ],
39
+ "mean": 0.19737919737919737,
40
+ "ci": 0.0
41
+ },
42
+ {
43
+ "task": "hellaswag",
44
+ "accuracies": [
45
+ 0.25731925911173076
46
+ ],
47
+ "mean": 0.25731925911173076,
48
+ "ci": 0.0
49
+ },
50
+ {
51
+ "task": "logiqa",
52
+ "accuracies": [
53
+ 0.2749615975422427
54
+ ],
55
+ "mean": 0.2749615975422427,
56
+ "ci": 0.0
57
+ },
58
+ {
59
+ "task": "mathqa",
60
+ "accuracies": [
61
+ 0.2254606365159129
62
+ ],
63
+ "mean": 0.2254606365159129,
64
+ "ci": 0.0
65
+ },
66
+ {
67
+ "task": "openbookqa",
68
+ "accuracies": [
69
+ 0.264
70
+ ],
71
+ "mean": 0.264,
72
+ "ci": 0.0
73
+ },
74
+ {
75
+ "task": "piqa",
76
+ "accuracies": [
77
+ 0.5174102285092492
78
+ ],
79
+ "mean": 0.5174102285092492,
80
+ "ci": 0.0
81
+ },
82
+ {
83
+ "task": "pubmedqa",
84
+ "accuracies": [
85
+ 0.452
86
+ ],
87
+ "mean": 0.452,
88
+ "ci": 0.0
89
+ },
90
+ {
91
+ "task": "winogrande",
92
+ "accuracies": [
93
+ 0.5232833464877664
94
+ ],
95
+ "mean": 0.5232833464877664,
96
+ "ci": 0.0
97
+ }
98
+ ],
99
+ "mean_accuracy": 0.3293628232525043
100
+ }
results/baseline-lr6e-4/checkpoint-3600.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_path": "/workspace/concept_model/runs/baseline-lr6e-4/checkpoint-3600",
3
+ "run": "baseline-lr6e-4",
4
+ "checkpoint": "checkpoint-3600",
5
+ "tasks": [
6
+ {
7
+ "task": "arc-challenge",
8
+ "accuracies": [
9
+ 0.2626609442060086
10
+ ],
11
+ "mean": 0.2626609442060086,
12
+ "ci": 0.0
13
+ },
14
+ {
15
+ "task": "arc-easy",
16
+ "accuracies": [
17
+ 0.2651162790697674
18
+ ],
19
+ "mean": 0.2651162790697674,
20
+ "ci": 0.0
21
+ },
22
+ {
23
+ "task": "boolq",
24
+ "accuracies": [
25
+ 0.6085626911314985
26
+ ],
27
+ "mean": 0.6085626911314985,
28
+ "ci": 0.0
29
+ },
30
+ {
31
+ "task": "commonsenseqa",
32
+ "accuracies": [
33
+ 0.20966420966420968
34
+ ],
35
+ "mean": 0.20966420966420968,
36
+ "ci": 0.0
37
+ },
38
+ {
39
+ "task": "hellaswag",
40
+ "accuracies": [
41
+ 0.25801633140808605
42
+ ],
43
+ "mean": 0.25801633140808605,
44
+ "ci": 0.0
45
+ },
46
+ {
47
+ "task": "logiqa",
48
+ "accuracies": [
49
+ 0.26881720430107525
50
+ ],
51
+ "mean": 0.26881720430107525,
52
+ "ci": 0.0
53
+ },
54
+ {
55
+ "task": "mathqa",
56
+ "accuracies": [
57
+ 0.21105527638190955
58
+ ],
59
+ "mean": 0.21105527638190955,
60
+ "ci": 0.0
61
+ },
62
+ {
63
+ "task": "openbookqa",
64
+ "accuracies": [
65
+ 0.268
66
+ ],
67
+ "mean": 0.268,
68
+ "ci": 0.0
69
+ },
70
+ {
71
+ "task": "piqa",
72
+ "accuracies": [
73
+ 0.4912948857453754
74
+ ],
75
+ "mean": 0.4912948857453754,
76
+ "ci": 0.0
77
+ },
78
+ {
79
+ "task": "pubmedqa",
80
+ "accuracies": [
81
+ 0.521
82
+ ],
83
+ "mean": 0.521,
84
+ "ci": 0.0
85
+ },
86
+ {
87
+ "task": "winogrande",
88
+ "accuracies": [
89
+ 0.48697711128650356
90
+ ],
91
+ "mean": 0.48697711128650356,
92
+ "ci": 0.0
93
+ }
94
+ ],
95
+ "mean_accuracy": 0.3501059030176758
96
+ }
results/baseline-lr6e-4/checkpoint-72.json ADDED
@@ -0,0 +1,118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_path": "/workspace/concept_model/runs/baseline-lr6e-4/checkpoint-72",
3
+ "run": "baseline-lr6e-4",
4
+ "checkpoint": "checkpoint-72",
5
+ "tasks": [
6
+ {
7
+ "task": "arc-challenge",
8
+ "accuracies": [
9
+ 0.2652360515021459,
10
+ 0.2532188841201717,
11
+ 0.2532188841201717
12
+ ],
13
+ "mean": 0.25722460658082974,
14
+ "ci": 0.017235232679969195
15
+ },
16
+ {
17
+ "task": "arc-easy",
18
+ "accuracies": [
19
+ 0.23594080338266385,
20
+ 0.23594080338266385,
21
+ 0.23594080338266385
22
+ ],
23
+ "mean": 0.23594080338266385,
24
+ "ci": NaN
25
+ },
26
+ {
27
+ "task": "boolq",
28
+ "accuracies": [
29
+ 0.6217125382262997,
30
+ 0.6217125382262997,
31
+ 0.6217125382262997
32
+ ],
33
+ "mean": 0.6217125382262997,
34
+ "ci": NaN
35
+ },
36
+ {
37
+ "task": "commonsenseqa",
38
+ "accuracies": [
39
+ 0.19983619983619982,
40
+ 0.19983619983619982,
41
+ 0.19983619983619982
42
+ ],
43
+ "mean": 0.19983619983619982,
44
+ "ci": NaN
45
+ },
46
+ {
47
+ "task": "hellaswag",
48
+ "accuracies": [
49
+ 0.2550288787094204,
50
+ 0.2550288787094204,
51
+ 0.2550288787094204
52
+ ],
53
+ "mean": 0.2550288787094204,
54
+ "ci": NaN
55
+ },
56
+ {
57
+ "task": "logiqa",
58
+ "accuracies": [
59
+ 0.2780337941628264,
60
+ 0.2780337941628264,
61
+ 0.2780337941628264
62
+ ],
63
+ "mean": 0.2780337941628264,
64
+ "ci": NaN
65
+ },
66
+ {
67
+ "task": "mathqa",
68
+ "accuracies": [
69
+ 0.22646566164154103,
70
+ 0.22646566164154103,
71
+ 0.22646566164154103
72
+ ],
73
+ "mean": 0.22646566164154103,
74
+ "ci": NaN
75
+ },
76
+ {
77
+ "task": "openbookqa",
78
+ "accuracies": [
79
+ 0.214,
80
+ 0.214,
81
+ 0.214
82
+ ],
83
+ "mean": 0.214,
84
+ "ci": NaN
85
+ },
86
+ {
87
+ "task": "piqa",
88
+ "accuracies": [
89
+ 0.5092491838955386,
90
+ 0.5092491838955386,
91
+ 0.5092491838955386
92
+ ],
93
+ "mean": 0.5092491838955386,
94
+ "ci": NaN
95
+ },
96
+ {
97
+ "task": "pubmedqa",
98
+ "accuracies": [
99
+ 0.127,
100
+ 0.127,
101
+ 0.127
102
+ ],
103
+ "mean": 0.127,
104
+ "ci": NaN
105
+ },
106
+ {
107
+ "task": "winogrande",
108
+ "accuracies": [
109
+ 0.5035516969218626,
110
+ 0.5035516969218626,
111
+ 0.5035516969218626
112
+ ],
113
+ "mean": 0.5035516969218626,
114
+ "ci": NaN
115
+ }
116
+ ],
117
+ "mean_accuracy": 0.31164030575974383
118
+ }
results/baseline-lr6e-4/checkpoint-720.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_path": "/workspace/concept_model/runs/baseline-lr6e-4/checkpoint-720",
3
+ "run": "baseline-lr6e-4",
4
+ "checkpoint": "checkpoint-720",
5
+ "tasks": [
6
+ {
7
+ "task": "arc-challenge",
8
+ "accuracies": [
9
+ 0.2575107296137339
10
+ ],
11
+ "mean": 0.2575107296137339,
12
+ "ci": 0.0
13
+ },
14
+ {
15
+ "task": "arc-easy",
16
+ "accuracies": [
17
+ 0.2693446088794926
18
+ ],
19
+ "mean": 0.2693446088794926,
20
+ "ci": 0.0
21
+ },
22
+ {
23
+ "task": "boolq",
24
+ "accuracies": [
25
+ 0.6201834862385321
26
+ ],
27
+ "mean": 0.6201834862385321,
28
+ "ci": 0.0
29
+ },
30
+ {
31
+ "task": "commonsenseqa",
32
+ "accuracies": [
33
+ 0.2022932022932023
34
+ ],
35
+ "mean": 0.2022932022932023,
36
+ "ci": 0.0
37
+ },
38
+ {
39
+ "task": "hellaswag",
40
+ "accuracies": [
41
+ 0.25941047600079664
42
+ ],
43
+ "mean": 0.25941047600079664,
44
+ "ci": 0.0
45
+ },
46
+ {
47
+ "task": "logiqa",
48
+ "accuracies": [
49
+ 0.2626728110599078
50
+ ],
51
+ "mean": 0.2626728110599078,
52
+ "ci": 0.0
53
+ },
54
+ {
55
+ "task": "mathqa",
56
+ "accuracies": [
57
+ 0.22144053601340033
58
+ ],
59
+ "mean": 0.22144053601340033,
60
+ "ci": 0.0
61
+ },
62
+ {
63
+ "task": "openbookqa",
64
+ "accuracies": [
65
+ 0.266
66
+ ],
67
+ "mean": 0.266,
68
+ "ci": 0.0
69
+ },
70
+ {
71
+ "task": "piqa",
72
+ "accuracies": [
73
+ 0.5021762785636561
74
+ ],
75
+ "mean": 0.5021762785636561,
76
+ "ci": 0.0
77
+ },
78
+ {
79
+ "task": "pubmedqa",
80
+ "accuracies": [
81
+ 0.552
82
+ ],
83
+ "mean": 0.552,
84
+ "ci": 0.0
85
+ },
86
+ {
87
+ "task": "winogrande",
88
+ "accuracies": [
89
+ 0.5074980268350434
90
+ ],
91
+ "mean": 0.5074980268350434,
92
+ "ci": 0.0
93
+ }
94
+ ],
95
+ "mean_accuracy": 0.3564118323179786
96
+ }