MaziyarPanahi
commited on
Commit
•
8fe6a22
1
Parent(s):
db8fde5
Update README.md (#1)
Browse files- Update README.md (cc29e273d00bd9e6f218ce2cf9e899f0ebf48a6e)
README.md
CHANGED
@@ -11,3 +11,387 @@ tags:
|
|
11 |
Merge of top 7B models with SLERP method.
|
12 |
|
13 |
> mergekit is a toolkit for merging pre-trained language models. mergekit uses an out-of-core approach to perform unreasonably elaborate merges in resource-constrained situations. Merges can be run entirely on CPU or accelerated with as little as 8 GB of VRAM. Many merging algorithms are supported, with more coming as they catch my attention.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
Merge of top 7B models with SLERP method.
|
12 |
|
13 |
> mergekit is a toolkit for merging pre-trained language models. mergekit uses an out-of-core approach to perform unreasonably elaborate merges in resource-constrained situations. Merges can be run entirely on CPU or accelerated with as little as 8 GB of VRAM. Many merging algorithms are supported, with more coming as they catch my attention.
|
14 |
+
>
|
15 |
+
> ## Eval
|
16 |
+
> ```python
|
17 |
+
> {
|
18 |
+
"all": {
|
19 |
+
"acc": 0.6545868511485138,
|
20 |
+
"acc_stderr": 0.031980293841566164,
|
21 |
+
"acc_norm": 0.6542757501692061,
|
22 |
+
"acc_norm_stderr": 0.03263807517879597,
|
23 |
+
"mc1": 0.45165238678090575,
|
24 |
+
"mc1_stderr": 0.017421480300277643,
|
25 |
+
"mc2": 0.6217500644350165,
|
26 |
+
"mc2_stderr": 0.015583825644663436
|
27 |
+
},
|
28 |
+
"harness|arc:challenge|25": {
|
29 |
+
"acc": 0.6723549488054608,
|
30 |
+
"acc_stderr": 0.01371584794071934,
|
31 |
+
"acc_norm": 0.6945392491467577,
|
32 |
+
"acc_norm_stderr": 0.01346008047800251
|
33 |
+
},
|
34 |
+
"harness|hellaswag|10": {
|
35 |
+
"acc": 0.7046405098585939,
|
36 |
+
"acc_stderr": 0.0045527183605131,
|
37 |
+
"acc_norm": 0.871539533957379,
|
38 |
+
"acc_norm_stderr": 0.0033391798350182853
|
39 |
+
},
|
40 |
+
"harness|hendrycksTest-abstract_algebra|5": {
|
41 |
+
"acc": 0.37,
|
42 |
+
"acc_stderr": 0.048523658709391,
|
43 |
+
"acc_norm": 0.37,
|
44 |
+
"acc_norm_stderr": 0.048523658709391
|
45 |
+
},
|
46 |
+
"harness|hendrycksTest-anatomy|5": {
|
47 |
+
"acc": 0.6148148148148148,
|
48 |
+
"acc_stderr": 0.04203921040156279,
|
49 |
+
"acc_norm": 0.6148148148148148,
|
50 |
+
"acc_norm_stderr": 0.04203921040156279
|
51 |
+
},
|
52 |
+
"harness|hendrycksTest-astronomy|5": {
|
53 |
+
"acc": 0.6907894736842105,
|
54 |
+
"acc_stderr": 0.037610708698674805,
|
55 |
+
"acc_norm": 0.6907894736842105,
|
56 |
+
"acc_norm_stderr": 0.037610708698674805
|
57 |
+
},
|
58 |
+
"harness|hendrycksTest-business_ethics|5": {
|
59 |
+
"acc": 0.63,
|
60 |
+
"acc_stderr": 0.04852365870939099,
|
61 |
+
"acc_norm": 0.63,
|
62 |
+
"acc_norm_stderr": 0.04852365870939099
|
63 |
+
},
|
64 |
+
"harness|hendrycksTest-clinical_knowledge|5": {
|
65 |
+
"acc": 0.7132075471698113,
|
66 |
+
"acc_stderr": 0.02783491252754407,
|
67 |
+
"acc_norm": 0.7132075471698113,
|
68 |
+
"acc_norm_stderr": 0.02783491252754407
|
69 |
+
},
|
70 |
+
"harness|hendrycksTest-college_biology|5": {
|
71 |
+
"acc": 0.7638888888888888,
|
72 |
+
"acc_stderr": 0.03551446610810826,
|
73 |
+
"acc_norm": 0.7638888888888888,
|
74 |
+
"acc_norm_stderr": 0.03551446610810826
|
75 |
+
},
|
76 |
+
"harness|hendrycksTest-college_chemistry|5": {
|
77 |
+
"acc": 0.47,
|
78 |
+
"acc_stderr": 0.050161355804659205,
|
79 |
+
"acc_norm": 0.47,
|
80 |
+
"acc_norm_stderr": 0.050161355804659205
|
81 |
+
},
|
82 |
+
"harness|hendrycksTest-college_computer_science|5": {
|
83 |
+
"acc": 0.57,
|
84 |
+
"acc_stderr": 0.04975698519562428,
|
85 |
+
"acc_norm": 0.57,
|
86 |
+
"acc_norm_stderr": 0.04975698519562428
|
87 |
+
},
|
88 |
+
"harness|hendrycksTest-college_mathematics|5": {
|
89 |
+
"acc": 0.27,
|
90 |
+
"acc_stderr": 0.0446196043338474,
|
91 |
+
"acc_norm": 0.27,
|
92 |
+
"acc_norm_stderr": 0.0446196043338474
|
93 |
+
},
|
94 |
+
"harness|hendrycksTest-college_medicine|5": {
|
95 |
+
"acc": 0.6820809248554913,
|
96 |
+
"acc_stderr": 0.0355068398916558,
|
97 |
+
"acc_norm": 0.6820809248554913,
|
98 |
+
"acc_norm_stderr": 0.0355068398916558
|
99 |
+
},
|
100 |
+
"harness|hendrycksTest-college_physics|5": {
|
101 |
+
"acc": 0.4019607843137255,
|
102 |
+
"acc_stderr": 0.04878608714466996,
|
103 |
+
"acc_norm": 0.4019607843137255,
|
104 |
+
"acc_norm_stderr": 0.04878608714466996
|
105 |
+
},
|
106 |
+
"harness|hendrycksTest-computer_security|5": {
|
107 |
+
"acc": 0.77,
|
108 |
+
"acc_stderr": 0.04229525846816506,
|
109 |
+
"acc_norm": 0.77,
|
110 |
+
"acc_norm_stderr": 0.04229525846816506
|
111 |
+
},
|
112 |
+
"harness|hendrycksTest-conceptual_physics|5": {
|
113 |
+
"acc": 0.5829787234042553,
|
114 |
+
"acc_stderr": 0.03223276266711712,
|
115 |
+
"acc_norm": 0.5829787234042553,
|
116 |
+
"acc_norm_stderr": 0.03223276266711712
|
117 |
+
},
|
118 |
+
"harness|hendrycksTest-econometrics|5": {
|
119 |
+
"acc": 0.5175438596491229,
|
120 |
+
"acc_stderr": 0.04700708033551038,
|
121 |
+
"acc_norm": 0.5175438596491229,
|
122 |
+
"acc_norm_stderr": 0.04700708033551038
|
123 |
+
},
|
124 |
+
"harness|hendrycksTest-electrical_engineering|5": {
|
125 |
+
"acc": 0.5586206896551724,
|
126 |
+
"acc_stderr": 0.04137931034482757,
|
127 |
+
"acc_norm": 0.5586206896551724,
|
128 |
+
"acc_norm_stderr": 0.04137931034482757
|
129 |
+
},
|
130 |
+
"harness|hendrycksTest-elementary_mathematics|5": {
|
131 |
+
"acc": 0.41798941798941797,
|
132 |
+
"acc_stderr": 0.025402555503260912,
|
133 |
+
"acc_norm": 0.41798941798941797,
|
134 |
+
"acc_norm_stderr": 0.025402555503260912
|
135 |
+
},
|
136 |
+
"harness|hendrycksTest-formal_logic|5": {
|
137 |
+
"acc": 0.47619047619047616,
|
138 |
+
"acc_stderr": 0.04467062628403273,
|
139 |
+
"acc_norm": 0.47619047619047616,
|
140 |
+
"acc_norm_stderr": 0.04467062628403273
|
141 |
+
},
|
142 |
+
"harness|hendrycksTest-global_facts|5": {
|
143 |
+
"acc": 0.34,
|
144 |
+
"acc_stderr": 0.04760952285695235,
|
145 |
+
"acc_norm": 0.34,
|
146 |
+
"acc_norm_stderr": 0.04760952285695235
|
147 |
+
},
|
148 |
+
"harness|hendrycksTest-high_school_biology|5": {
|
149 |
+
"acc": 0.7903225806451613,
|
150 |
+
"acc_stderr": 0.023157879349083522,
|
151 |
+
"acc_norm": 0.7903225806451613,
|
152 |
+
"acc_norm_stderr": 0.023157879349083522
|
153 |
+
},
|
154 |
+
"harness|hendrycksTest-high_school_chemistry|5": {
|
155 |
+
"acc": 0.4975369458128079,
|
156 |
+
"acc_stderr": 0.03517945038691063,
|
157 |
+
"acc_norm": 0.4975369458128079,
|
158 |
+
"acc_norm_stderr": 0.03517945038691063
|
159 |
+
},
|
160 |
+
"harness|hendrycksTest-high_school_computer_science|5": {
|
161 |
+
"acc": 0.68,
|
162 |
+
"acc_stderr": 0.04688261722621505,
|
163 |
+
"acc_norm": 0.68,
|
164 |
+
"acc_norm_stderr": 0.04688261722621505
|
165 |
+
},
|
166 |
+
"harness|hendrycksTest-high_school_european_history|5": {
|
167 |
+
"acc": 0.7696969696969697,
|
168 |
+
"acc_stderr": 0.0328766675860349,
|
169 |
+
"acc_norm": 0.7696969696969697,
|
170 |
+
"acc_norm_stderr": 0.0328766675860349
|
171 |
+
},
|
172 |
+
"harness|hendrycksTest-high_school_geography|5": {
|
173 |
+
"acc": 0.7828282828282829,
|
174 |
+
"acc_stderr": 0.029376616484945633,
|
175 |
+
"acc_norm": 0.7828282828282829,
|
176 |
+
"acc_norm_stderr": 0.029376616484945633
|
177 |
+
},
|
178 |
+
"harness|hendrycksTest-high_school_government_and_politics|5": {
|
179 |
+
"acc": 0.9015544041450777,
|
180 |
+
"acc_stderr": 0.021500249576033456,
|
181 |
+
"acc_norm": 0.9015544041450777,
|
182 |
+
"acc_norm_stderr": 0.021500249576033456
|
183 |
+
},
|
184 |
+
"harness|hendrycksTest-high_school_macroeconomics|5": {
|
185 |
+
"acc": 0.6717948717948717,
|
186 |
+
"acc_stderr": 0.023807633198657266,
|
187 |
+
"acc_norm": 0.6717948717948717,
|
188 |
+
"acc_norm_stderr": 0.023807633198657266
|
189 |
+
},
|
190 |
+
"harness|hendrycksTest-high_school_mathematics|5": {
|
191 |
+
"acc": 0.34444444444444444,
|
192 |
+
"acc_stderr": 0.02897264888484427,
|
193 |
+
"acc_norm": 0.34444444444444444,
|
194 |
+
"acc_norm_stderr": 0.02897264888484427
|
195 |
+
},
|
196 |
+
"harness|hendrycksTest-high_school_microeconomics|5": {
|
197 |
+
"acc": 0.6638655462184874,
|
198 |
+
"acc_stderr": 0.030684737115135363,
|
199 |
+
"acc_norm": 0.6638655462184874,
|
200 |
+
"acc_norm_stderr": 0.030684737115135363
|
201 |
+
},
|
202 |
+
"harness|hendrycksTest-high_school_physics|5": {
|
203 |
+
"acc": 0.304635761589404,
|
204 |
+
"acc_stderr": 0.03757949922943343,
|
205 |
+
"acc_norm": 0.304635761589404,
|
206 |
+
"acc_norm_stderr": 0.03757949922943343
|
207 |
+
},
|
208 |
+
"harness|hendrycksTest-high_school_psychology|5": {
|
209 |
+
"acc": 0.8458715596330275,
|
210 |
+
"acc_stderr": 0.015480826865374303,
|
211 |
+
"acc_norm": 0.8458715596330275,
|
212 |
+
"acc_norm_stderr": 0.015480826865374303
|
213 |
+
},
|
214 |
+
"harness|hendrycksTest-high_school_statistics|5": {
|
215 |
+
"acc": 0.5185185185185185,
|
216 |
+
"acc_stderr": 0.03407632093854051,
|
217 |
+
"acc_norm": 0.5185185185185185,
|
218 |
+
"acc_norm_stderr": 0.03407632093854051
|
219 |
+
},
|
220 |
+
"harness|hendrycksTest-high_school_us_history|5": {
|
221 |
+
"acc": 0.8382352941176471,
|
222 |
+
"acc_stderr": 0.025845017986926917,
|
223 |
+
"acc_norm": 0.8382352941176471,
|
224 |
+
"acc_norm_stderr": 0.025845017986926917
|
225 |
+
},
|
226 |
+
"harness|hendrycksTest-high_school_world_history|5": {
|
227 |
+
"acc": 0.810126582278481,
|
228 |
+
"acc_stderr": 0.02553010046023349,
|
229 |
+
"acc_norm": 0.810126582278481,
|
230 |
+
"acc_norm_stderr": 0.02553010046023349
|
231 |
+
},
|
232 |
+
"harness|hendrycksTest-human_aging|5": {
|
233 |
+
"acc": 0.6905829596412556,
|
234 |
+
"acc_stderr": 0.03102441174057221,
|
235 |
+
"acc_norm": 0.6905829596412556,
|
236 |
+
"acc_norm_stderr": 0.03102441174057221
|
237 |
+
},
|
238 |
+
"harness|hendrycksTest-human_sexuality|5": {
|
239 |
+
"acc": 0.7786259541984732,
|
240 |
+
"acc_stderr": 0.036412970813137296,
|
241 |
+
"acc_norm": 0.7786259541984732,
|
242 |
+
"acc_norm_stderr": 0.036412970813137296
|
243 |
+
},
|
244 |
+
"harness|hendrycksTest-international_law|5": {
|
245 |
+
"acc": 0.8099173553719008,
|
246 |
+
"acc_stderr": 0.03581796951709282,
|
247 |
+
"acc_norm": 0.8099173553719008,
|
248 |
+
"acc_norm_stderr": 0.03581796951709282
|
249 |
+
},
|
250 |
+
"harness|hendrycksTest-jurisprudence|5": {
|
251 |
+
"acc": 0.7685185185185185,
|
252 |
+
"acc_stderr": 0.04077494709252627,
|
253 |
+
"acc_norm": 0.7685185185185185,
|
254 |
+
"acc_norm_stderr": 0.04077494709252627
|
255 |
+
},
|
256 |
+
"harness|hendrycksTest-logical_fallacies|5": {
|
257 |
+
"acc": 0.7607361963190185,
|
258 |
+
"acc_stderr": 0.0335195387952127,
|
259 |
+
"acc_norm": 0.7607361963190185,
|
260 |
+
"acc_norm_stderr": 0.0335195387952127
|
261 |
+
},
|
262 |
+
"harness|hendrycksTest-machine_learning|5": {
|
263 |
+
"acc": 0.45535714285714285,
|
264 |
+
"acc_stderr": 0.047268355537191,
|
265 |
+
"acc_norm": 0.45535714285714285,
|
266 |
+
"acc_norm_stderr": 0.047268355537191
|
267 |
+
},
|
268 |
+
"harness|hendrycksTest-management|5": {
|
269 |
+
"acc": 0.8058252427184466,
|
270 |
+
"acc_stderr": 0.03916667762822584,
|
271 |
+
"acc_norm": 0.8058252427184466,
|
272 |
+
"acc_norm_stderr": 0.03916667762822584
|
273 |
+
},
|
274 |
+
"harness|hendrycksTest-marketing|5": {
|
275 |
+
"acc": 0.8675213675213675,
|
276 |
+
"acc_stderr": 0.022209309073165612,
|
277 |
+
"acc_norm": 0.8675213675213675,
|
278 |
+
"acc_norm_stderr": 0.022209309073165612
|
279 |
+
},
|
280 |
+
"harness|hendrycksTest-medical_genetics|5": {
|
281 |
+
"acc": 0.71,
|
282 |
+
"acc_stderr": 0.045604802157206845,
|
283 |
+
"acc_norm": 0.71,
|
284 |
+
"acc_norm_stderr": 0.045604802157206845
|
285 |
+
},
|
286 |
+
"harness|hendrycksTest-miscellaneous|5": {
|
287 |
+
"acc": 0.8352490421455939,
|
288 |
+
"acc_stderr": 0.013265346261323788,
|
289 |
+
"acc_norm": 0.8352490421455939,
|
290 |
+
"acc_norm_stderr": 0.013265346261323788
|
291 |
+
},
|
292 |
+
"harness|hendrycksTest-moral_disputes|5": {
|
293 |
+
"acc": 0.7543352601156069,
|
294 |
+
"acc_stderr": 0.023176298203992005,
|
295 |
+
"acc_norm": 0.7543352601156069,
|
296 |
+
"acc_norm_stderr": 0.023176298203992005
|
297 |
+
},
|
298 |
+
"harness|hendrycksTest-moral_scenarios|5": {
|
299 |
+
"acc": 0.4547486033519553,
|
300 |
+
"acc_stderr": 0.016653875777524006,
|
301 |
+
"acc_norm": 0.4547486033519553,
|
302 |
+
"acc_norm_stderr": 0.016653875777524006
|
303 |
+
},
|
304 |
+
"harness|hendrycksTest-nutrition|5": {
|
305 |
+
"acc": 0.7483660130718954,
|
306 |
+
"acc_stderr": 0.0248480182638752,
|
307 |
+
"acc_norm": 0.7483660130718954,
|
308 |
+
"acc_norm_stderr": 0.0248480182638752
|
309 |
+
},
|
310 |
+
"harness|hendrycksTest-philosophy|5": {
|
311 |
+
"acc": 0.7202572347266881,
|
312 |
+
"acc_stderr": 0.02549425935069491,
|
313 |
+
"acc_norm": 0.7202572347266881,
|
314 |
+
"acc_norm_stderr": 0.02549425935069491
|
315 |
+
},
|
316 |
+
"harness|hendrycksTest-prehistory|5": {
|
317 |
+
"acc": 0.7592592592592593,
|
318 |
+
"acc_stderr": 0.02378858355165854,
|
319 |
+
"acc_norm": 0.7592592592592593,
|
320 |
+
"acc_norm_stderr": 0.02378858355165854
|
321 |
+
},
|
322 |
+
"harness|hendrycksTest-professional_accounting|5": {
|
323 |
+
"acc": 0.4787234042553192,
|
324 |
+
"acc_stderr": 0.029800481645628693,
|
325 |
+
"acc_norm": 0.4787234042553192,
|
326 |
+
"acc_norm_stderr": 0.029800481645628693
|
327 |
+
},
|
328 |
+
"harness|hendrycksTest-professional_law|5": {
|
329 |
+
"acc": 0.4745762711864407,
|
330 |
+
"acc_stderr": 0.012753716929101008,
|
331 |
+
"acc_norm": 0.4745762711864407,
|
332 |
+
"acc_norm_stderr": 0.012753716929101008
|
333 |
+
},
|
334 |
+
"harness|hendrycksTest-professional_medicine|5": {
|
335 |
+
"acc": 0.7095588235294118,
|
336 |
+
"acc_stderr": 0.027576468622740536,
|
337 |
+
"acc_norm": 0.7095588235294118,
|
338 |
+
"acc_norm_stderr": 0.027576468622740536
|
339 |
+
},
|
340 |
+
"harness|hendrycksTest-professional_psychology|5": {
|
341 |
+
"acc": 0.6928104575163399,
|
342 |
+
"acc_stderr": 0.01866335967146367,
|
343 |
+
"acc_norm": 0.6928104575163399,
|
344 |
+
"acc_norm_stderr": 0.01866335967146367
|
345 |
+
},
|
346 |
+
"harness|hendrycksTest-public_relations|5": {
|
347 |
+
"acc": 0.6727272727272727,
|
348 |
+
"acc_stderr": 0.0449429086625209,
|
349 |
+
"acc_norm": 0.6727272727272727,
|
350 |
+
"acc_norm_stderr": 0.0449429086625209
|
351 |
+
},
|
352 |
+
"harness|hendrycksTest-security_studies|5": {
|
353 |
+
"acc": 0.7387755102040816,
|
354 |
+
"acc_stderr": 0.02812342933514278,
|
355 |
+
"acc_norm": 0.7387755102040816,
|
356 |
+
"acc_norm_stderr": 0.02812342933514278
|
357 |
+
},
|
358 |
+
"harness|hendrycksTest-sociology|5": {
|
359 |
+
"acc": 0.845771144278607,
|
360 |
+
"acc_stderr": 0.025538433368578337,
|
361 |
+
"acc_norm": 0.845771144278607,
|
362 |
+
"acc_norm_stderr": 0.025538433368578337
|
363 |
+
},
|
364 |
+
"harness|hendrycksTest-us_foreign_policy|5": {
|
365 |
+
"acc": 0.86,
|
366 |
+
"acc_stderr": 0.0348735088019777,
|
367 |
+
"acc_norm": 0.86,
|
368 |
+
"acc_norm_stderr": 0.0348735088019777
|
369 |
+
},
|
370 |
+
"harness|hendrycksTest-virology|5": {
|
371 |
+
"acc": 0.5481927710843374,
|
372 |
+
"acc_stderr": 0.03874371556587953,
|
373 |
+
"acc_norm": 0.5481927710843374,
|
374 |
+
"acc_norm_stderr": 0.03874371556587953
|
375 |
+
},
|
376 |
+
"harness|hendrycksTest-world_religions|5": {
|
377 |
+
"acc": 0.8362573099415205,
|
378 |
+
"acc_stderr": 0.028380919596145866,
|
379 |
+
"acc_norm": 0.8362573099415205,
|
380 |
+
"acc_norm_stderr": 0.028380919596145866
|
381 |
+
},
|
382 |
+
"harness|truthfulqa:mc|0": {
|
383 |
+
"mc1": 0.45165238678090575,
|
384 |
+
"mc1_stderr": 0.017421480300277643,
|
385 |
+
"mc2": 0.6217500644350165,
|
386 |
+
"mc2_stderr": 0.015583825644663436
|
387 |
+
},
|
388 |
+
"harness|winogrande|5": {
|
389 |
+
"acc": 0.7963693764798737,
|
390 |
+
"acc_stderr": 0.011317798781626913
|
391 |
+
},
|
392 |
+
"harness|gsm8k|5": {
|
393 |
+
"acc": 0.7202426080363912,
|
394 |
+
"acc_stderr": 0.01236438401673532
|
395 |
+
}
|
396 |
+
}
|
397 |
+
```
|