Muennighoff commited on
Commit
c947298
1 Parent(s): f3dc0fe
4b284bc4dedup/evaluation/rankeval/4b284bc4dedup_5.csv CHANGED
@@ -12,6 +12,10 @@ cb,f1,0.2980010185892538,,1
12
  copa,acc,0.81,0.03942772444036623,0
13
  hellaswag,acc,0.4772953594901414,0.004984634285101618,0
14
  hellaswag,acc_norm,0.6417048396733719,0.00478519504988916,0
 
 
15
  rte,acc,0.5126353790613718,0.030086851767188564,0
 
 
16
  storycloze_2016,acc,0.7439871726349546,0.01009236116036427,0
17
  winogrande,acc,0.5974743488555643,0.013782866831703048,0
 
12
  copa,acc,0.81,0.03942772444036623,0
13
  hellaswag,acc,0.4772953594901414,0.004984634285101618,0
14
  hellaswag,acc_norm,0.6417048396733719,0.00478519504988916,0
15
+ piqa,acc,0.7698585418933623,0.00982083282683981,0
16
+ piqa,acc_norm,0.779107725788901,0.009679088048842207,0
17
  rte,acc,0.5126353790613718,0.030086851767188564,0
18
+ sciq,acc,0.887,0.01001655286669685,0
19
+ sciq,acc_norm,0.876,0.01042749887234396,0
20
  storycloze_2016,acc,0.7439871726349546,0.01009236116036427,0
21
  winogrande,acc,0.5974743488555643,0.013782866831703048,0
4b284bc4dedup/evaluation/rankeval/4b284bc4dedup_5.json CHANGED
@@ -54,6 +54,18 @@
54
  "acc_stderr": 0.013250012579393443,
55
  "acc_norm": 0.30716723549488056,
56
  "acc_norm_stderr": 0.013481034054980945
 
 
 
 
 
 
 
 
 
 
 
 
57
  }
58
  },
59
  "versions": {
@@ -68,6 +80,8 @@
68
  "storycloze_2016": 0,
69
  "boolq": 1,
70
  "arc_easy": 0,
71
- "arc_challenge": 0
 
 
72
  }
73
  }
 
54
  "acc_stderr": 0.013250012579393443,
55
  "acc_norm": 0.30716723549488056,
56
  "acc_norm_stderr": 0.013481034054980945
57
+ },
58
+ "sciq": {
59
+ "acc": 0.887,
60
+ "acc_stderr": 0.01001655286669685,
61
+ "acc_norm": 0.876,
62
+ "acc_norm_stderr": 0.01042749887234396
63
+ },
64
+ "piqa": {
65
+ "acc": 0.7698585418933623,
66
+ "acc_stderr": 0.00982083282683981,
67
+ "acc_norm": 0.779107725788901,
68
+ "acc_norm_stderr": 0.009679088048842207
69
  }
70
  },
71
  "versions": {
 
80
  "storycloze_2016": 0,
81
  "boolq": 1,
82
  "arc_easy": 0,
83
+ "arc_challenge": 0,
84
+ "sciq": 0,
85
+ "piqa": 0
86
  }
87
  }