Hugging Face's logo

AbstractPhil
/
geolip-rotary-experiments

Safetensors
rope
rotary-position-embedding
positional-encoding
mixture-of-experts
model-merging
associative-recall
experiments
Model card Files
xet
 Community
AbstractPhil commited on
Commit
c47df29
·
verified ·
1 Parent(s): 6b840b3

scale members: alhena-both_s5_L48000_ctx4096_frdyn_cap_mode_sz32-128-256-512_r4_n30_o0_askany_p75d_gag50_cp4_gc48_ac44_kh50_va1_pt10b4h0_krt16r218_ahdigit1_a48long_look_b4h0

Browse files
Files changed (1) hide show
  1. bed/reports/look/alhena-both_s5_L48000_ctx4096_frdyn_cap_mode_sz32-128-256-512_r4_n30_o0_askany_p75d_gag50_cp4_gc48_ac44_kh50_va1_pt10b4h0_krt16r218_ahdigit1_a48long_look_b4h0.json +195 -0
bed/reports/look/alhena-both_s5_L48000_ctx4096_frdyn_cap_mode_sz32-128-256-512_r4_n30_o0_askany_p75d_gag50_cp4_gc48_ac44_kh50_va1_pt10b4h0_krt16r218_ahdigit1_a48long_look_b4h0.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "core": "alhena-both_s5_L48000_ctx4096_frdyn_cap_mode_sz32-128-256-512_r4_n30_o0_askany_p75d_gag50_cp4_gc48_ac44_kh50_va1_pt10b4h0_krt16r218_ahdigit1_a48long",
3
+ "seed": 5,
4
+ "version": "0.1.100",
5
+ "look": {
6
+ "block": 4,
7
+ "head": 0,
8
+ "thr": 0.5,
9
+ "from_ptr": true
10
+ },
11
+ "sets": {
12
+ "read": 512,
13
+ "hard": 128
14
+ },
15
+ "views": [
16
+ "LAYOUT",
17
+ "INDEXF",
18
+ "LOOKF",
19
+ "LOOKQF"
20
+ ],
21
+ "requests": [
22
+ "INDEX",
23
+ "INDEXF",
24
+ "LOOK",
25
+ "LOOKF",
26
+ "LOOKQF"
27
+ ],
28
+ "read": {
29
+ "rows": 512,
30
+ "questions": 12288,
31
+ "request": {
32
+ "INDEX": {
33
+ "right": 0.8068033854166666,
34
+ "no_request": 0.0,
35
+ "wrong": 0.19319661458333337
36
+ },
37
+ "INDEXF": {
38
+ "right": 0.7562662760416666,
39
+ "no_request": 0.0,
40
+ "wrong": 0.24373372395833337
41
+ },
42
+ "LOOK": {
43
+ "right": 1.0,
44
+ "no_request": 0.0,
45
+ "wrong": 0.0
46
+ },
47
+ "LOOKF": {
48
+ "right": 1.0,
49
+ "no_request": 0.0,
50
+ "wrong": 0.0
51
+ },
52
+ "LOOKQF": {
53
+ "right": 1.0,
54
+ "no_request": 0.0,
55
+ "wrong": 0.0
56
+ }
57
+ },
58
+ "answers": {
59
+ "LAYOUT": 0.9989420572916666,
60
+ "INDEXF": 0.75830078125,
61
+ "LOOKF": 0.9989420572916666,
62
+ "LOOKQF": 0.9989420572916666
63
+ },
64
+ "look": {
65
+ "LOOK": {
66
+ "questions": 12288,
67
+ "right": 1.0,
68
+ "no_request": 0.0,
69
+ "wrong_record": 0.0,
70
+ "decoded_mass_mean": 0.9999950906026114,
71
+ "true_mass_mean": 0.9999950906026114,
72
+ "largest_mass_form_right": 1.0,
73
+ "split": 0.0,
74
+ "duplicate": 0.0,
75
+ "wrong_same_ending": 0.0,
76
+ "wrong_records": 0
77
+ },
78
+ "LOOKF": {
79
+ "questions": 12288,
80
+ "right": 1.0,
81
+ "no_request": 0.0,
82
+ "wrong_record": 0.0,
83
+ "decoded_mass_mean": 0.9999950892733654,
84
+ "true_mass_mean": 0.9999950892733654,
85
+ "largest_mass_form_right": 1.0,
86
+ "split": 0.0,
87
+ "duplicate": 0.0,
88
+ "wrong_same_ending": 0.0,
89
+ "wrong_records": 0
90
+ },
91
+ "LOOKQF": {
92
+ "questions": 12288,
93
+ "right": 1.0,
94
+ "no_request": 0.0,
95
+ "wrong_record": 0.0,
96
+ "decoded_mass_mean": 0.9999941362517277,
97
+ "true_mass_mean": 0.9999941362517277,
98
+ "largest_mass_form_right": 1.0,
99
+ "split": 0.0,
100
+ "duplicate": 0.0,
101
+ "wrong_same_ending": 0.0,
102
+ "wrong_records": 0
103
+ }
104
+ },
105
+ "free_running": {
106
+ "passes_mean": 3.78125,
107
+ "passes_max": 5,
108
+ "converged": 1.0
109
+ }
110
+ },
111
+ "hard": {
112
+ "rows": 128,
113
+ "questions": 3072,
114
+ "request": {
115
+ "INDEX": {
116
+ "right": 0.8059895833333334,
117
+ "no_request": 0.0,
118
+ "wrong": 0.19401041666666663
119
+ },
120
+ "INDEXF": {
121
+ "right": 0.7529296875,
122
+ "no_request": 0.0,
123
+ "wrong": 0.2470703125
124
+ },
125
+ "LOOK": {
126
+ "right": 1.0,
127
+ "no_request": 0.0,
128
+ "wrong": 0.0
129
+ },
130
+ "LOOKF": {
131
+ "right": 1.0,
132
+ "no_request": 0.0,
133
+ "wrong": 0.0
134
+ },
135
+ "LOOKQF": {
136
+ "right": 1.0,
137
+ "no_request": 0.0,
138
+ "wrong": 0.0
139
+ }
140
+ },
141
+ "answers": {
142
+ "LAYOUT": 0.9986979166666666,
143
+ "INDEXF": 0.7526041666666666,
144
+ "LOOKF": 0.9986979166666666,
145
+ "LOOKQF": 0.9986979166666666
146
+ },
147
+ "look": {
148
+ "LOOK": {
149
+ "questions": 3072,
150
+ "right": 1.0,
151
+ "no_request": 0.0,
152
+ "wrong_record": 0.0,
153
+ "decoded_mass_mean": 0.9999987433718566,
154
+ "true_mass_mean": 0.9999987433718566,
155
+ "largest_mass_form_right": 1.0,
156
+ "split": 0.0,
157
+ "duplicate": 0.0,
158
+ "wrong_same_ending": 0.0,
159
+ "wrong_records": 0
160
+ },
161
+ "LOOKF": {
162
+ "questions": 3072,
163
+ "right": 1.0,
164
+ "no_request": 0.0,
165
+ "wrong_record": 0.0,
166
+ "decoded_mass_mean": 0.9999987432167058,
167
+ "true_mass_mean": 0.9999987432167058,
168
+ "largest_mass_form_right": 1.0,
169
+ "split": 0.0,
170
+ "duplicate": 0.0,
171
+ "wrong_same_ending": 0.0,
172
+ "wrong_records": 0
173
+ },
174
+ "LOOKQF": {
175
+ "questions": 3072,
176
+ "right": 1.0,
177
+ "no_request": 0.0,
178
+ "wrong_record": 0.0,
179
+ "decoded_mass_mean": 0.9999978891115341,
180
+ "true_mass_mean": 0.9999978891115341,
181
+ "largest_mass_form_right": 1.0,
182
+ "split": 0.0,
183
+ "duplicate": 0.0,
184
+ "wrong_same_ending": 0.0,
185
+ "wrong_records": 0
186
+ }
187
+ },
188
+ "free_running": {
189
+ "passes_mean": 3.6875,
190
+ "passes_max": 5,
191
+ "converged": 1.0
192
+ }
193
+ },
194
+ "wall_s": 30.069814205169678
195
+ }