Delete mergekit_config.yml
Browse files- mergekit_config.yml +0 -434
mergekit_config.yml
DELETED
@@ -1,434 +0,0 @@
|
|
1 |
-
# Six splits plus "end game
|
2 |
-
# "D" starts at plus .1 VS D/O proj.
|
3 |
-
# 40 plus.
|
4 |
-
|
5 |
-
slices:
|
6 |
-
- sources:
|
7 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
8 |
-
layer_range: [0, 27]
|
9 |
-
|
10 |
-
# conc layers
|
11 |
-
# split 1
|
12 |
-
|
13 |
-
- sources:
|
14 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
15 |
-
layer_range: [27,28]
|
16 |
-
parameters:
|
17 |
-
scale:
|
18 |
-
- filter: o_proj
|
19 |
-
value: 0.01
|
20 |
-
- filter: down_proj
|
21 |
-
value: 0.01
|
22 |
-
- value: 0.11
|
23 |
-
- sources:
|
24 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
25 |
-
layer_range: [27,28]
|
26 |
-
parameters:
|
27 |
-
scale:
|
28 |
-
- filter: o_proj
|
29 |
-
value: 0.02
|
30 |
-
- filter: down_proj
|
31 |
-
value: 0.02
|
32 |
-
- value: 0.12
|
33 |
-
- sources:
|
34 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
35 |
-
layer_range: [27,28]
|
36 |
-
parameters:
|
37 |
-
scale:
|
38 |
-
- filter: o_proj
|
39 |
-
value: 0.03
|
40 |
-
- filter: down_proj
|
41 |
-
value: 0.03
|
42 |
-
- value: 0.13
|
43 |
-
|
44 |
-
- sources:
|
45 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
46 |
-
layer_range: [27,28]
|
47 |
-
parameters:
|
48 |
-
scale:
|
49 |
-
- filter: o_proj
|
50 |
-
value: 0.04
|
51 |
-
- filter: down_proj
|
52 |
-
value: 0.04
|
53 |
-
- value: 0.61
|
54 |
-
|
55 |
-
# split 2, SURGE D THEN D drop .46, continues @ D .15 (from .13)
|
56 |
-
|
57 |
-
- sources:
|
58 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
59 |
-
layer_range: [27,28]
|
60 |
-
parameters:
|
61 |
-
scale:
|
62 |
-
- filter: o_proj
|
63 |
-
value: 0.05
|
64 |
-
- filter: down_proj
|
65 |
-
value: 0.05
|
66 |
-
- value: 0.15
|
67 |
-
- sources:
|
68 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
69 |
-
layer_range: [27,28]
|
70 |
-
parameters:
|
71 |
-
scale:
|
72 |
-
- filter: o_proj
|
73 |
-
value: 0.06
|
74 |
-
- filter: down_proj
|
75 |
-
value: 0.06
|
76 |
-
- value: 0.16
|
77 |
-
- sources:
|
78 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
79 |
-
layer_range: [27,28]
|
80 |
-
parameters:
|
81 |
-
scale:
|
82 |
-
- filter: o_proj
|
83 |
-
value: 0.07
|
84 |
-
- filter: down_proj
|
85 |
-
value: 0.07
|
86 |
-
- value: 0.17
|
87 |
-
- sources:
|
88 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
89 |
-
layer_range: [27,28]
|
90 |
-
parameters:
|
91 |
-
scale:
|
92 |
-
- filter: o_proj
|
93 |
-
value: 0.08
|
94 |
-
- filter: down_proj
|
95 |
-
value: 0.08
|
96 |
-
- value: 0.41
|
97 |
-
|
98 |
-
# split 3, SURGE D to .41, D drop .21 ... follows .17 previous
|
99 |
-
|
100 |
-
- sources:
|
101 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
102 |
-
layer_range: [27,28]
|
103 |
-
parameters:
|
104 |
-
scale:
|
105 |
-
- filter: o_proj
|
106 |
-
value: 0.09
|
107 |
-
- filter: down_proj
|
108 |
-
value: 0.09
|
109 |
-
- value: 0.19
|
110 |
-
- sources:
|
111 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
112 |
-
layer_range: [27,28]
|
113 |
-
parameters:
|
114 |
-
scale:
|
115 |
-
- filter: o_proj
|
116 |
-
value: 0.10
|
117 |
-
- filter: down_proj
|
118 |
-
value: 0.10
|
119 |
-
- value: 0.20
|
120 |
-
- sources:
|
121 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
122 |
-
layer_range: [27,28]
|
123 |
-
parameters:
|
124 |
-
scale:
|
125 |
-
- filter: o_proj
|
126 |
-
value: 0.11
|
127 |
-
- filter: down_proj
|
128 |
-
value: 0.11
|
129 |
-
- value: .22
|
130 |
-
- sources:
|
131 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
132 |
-
layer_range: [27,28]
|
133 |
-
parameters:
|
134 |
-
scale:
|
135 |
-
- filter: o_proj
|
136 |
-
value: 0.12
|
137 |
-
- filter: down_proj
|
138 |
-
value: 0.12
|
139 |
-
- value: .24
|
140 |
-
- sources:
|
141 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
142 |
-
layer_range: [27,28]
|
143 |
-
parameters:
|
144 |
-
scale:
|
145 |
-
- filter: o_proj
|
146 |
-
value: 0.13
|
147 |
-
- filter: down_proj
|
148 |
-
value: 0.13
|
149 |
-
- value: .26
|
150 |
-
- sources:
|
151 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
152 |
-
layer_range: [27,28]
|
153 |
-
parameters:
|
154 |
-
scale:
|
155 |
-
- filter: o_proj
|
156 |
-
value: 0.14
|
157 |
-
- filter: down_proj
|
158 |
-
value: 0.14
|
159 |
-
- value: .28
|
160 |
-
- sources:
|
161 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
162 |
-
layer_range: [27,28]
|
163 |
-
parameters:
|
164 |
-
scale:
|
165 |
-
- filter: o_proj
|
166 |
-
value: 0.15
|
167 |
-
- filter: down_proj
|
168 |
-
value: 0.15
|
169 |
-
- value: .30
|
170 |
-
- sources:
|
171 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
172 |
-
layer_range: [27,28]
|
173 |
-
parameters:
|
174 |
-
scale:
|
175 |
-
- filter: o_proj
|
176 |
-
value: 0.16
|
177 |
-
- filter: down_proj
|
178 |
-
value: 0.16
|
179 |
-
- value: .31
|
180 |
-
- sources:
|
181 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
182 |
-
layer_range: [27,28]
|
183 |
-
parameters:
|
184 |
-
scale:
|
185 |
-
- filter: o_proj
|
186 |
-
value: 0.20
|
187 |
-
- filter: down_proj
|
188 |
-
value: 0.20
|
189 |
-
- value: .32
|
190 |
-
- sources:
|
191 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
192 |
-
layer_range: [27,28]
|
193 |
-
parameters:
|
194 |
-
scale:
|
195 |
-
- filter: o_proj
|
196 |
-
value: 0.21
|
197 |
-
- filter: down_proj
|
198 |
-
value: 0.21
|
199 |
-
- value: .33
|
200 |
-
- sources:
|
201 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
202 |
-
layer_range: [27,28]
|
203 |
-
parameters:
|
204 |
-
scale:
|
205 |
-
- filter: o_proj
|
206 |
-
value: 0.22
|
207 |
-
- filter: down_proj
|
208 |
-
value: 0.22
|
209 |
-
- value: .34
|
210 |
-
- sources:
|
211 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
212 |
-
layer_range: [27,28]
|
213 |
-
parameters:
|
214 |
-
scale:
|
215 |
-
- filter: o_proj
|
216 |
-
value: 0.23
|
217 |
-
- filter: down_proj
|
218 |
-
value: 0.23
|
219 |
-
- value: .35
|
220 |
-
|
221 |
-
# split 4 , NO SURGE D, "D" down drop of .24 ; reverts to .11 (the very first "D" setting )
|
222 |
-
|
223 |
-
- sources:
|
224 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
225 |
-
layer_range: [27,28]
|
226 |
-
parameters:
|
227 |
-
scale:
|
228 |
-
- filter: o_proj
|
229 |
-
value: 0.24
|
230 |
-
- filter: down_proj
|
231 |
-
value: 0.24
|
232 |
-
- value: 0.11
|
233 |
-
- sources:
|
234 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
235 |
-
layer_range: [27,28]
|
236 |
-
parameters:
|
237 |
-
scale:
|
238 |
-
- filter: o_proj
|
239 |
-
value: 0.241
|
240 |
-
- filter: down_proj
|
241 |
-
value: 0.241
|
242 |
-
- value: 0.12
|
243 |
-
- sources:
|
244 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
245 |
-
layer_range: [27,28]
|
246 |
-
parameters:
|
247 |
-
scale:
|
248 |
-
- filter: o_proj
|
249 |
-
value: 0.242
|
250 |
-
- filter: down_proj
|
251 |
-
value: 0.243
|
252 |
-
- value: 0.13
|
253 |
-
- sources:
|
254 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
255 |
-
layer_range: [27,28]
|
256 |
-
parameters:
|
257 |
-
scale:
|
258 |
-
- filter: o_proj
|
259 |
-
value: 0.244
|
260 |
-
- filter: down_proj
|
261 |
-
value: 0.244
|
262 |
-
- value: 0.61
|
263 |
-
|
264 |
-
# split 5, D Surge to .61, drop to .15 (following .13)
|
265 |
-
|
266 |
-
- sources:
|
267 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
268 |
-
layer_range: [27,28]
|
269 |
-
parameters:
|
270 |
-
scale:
|
271 |
-
- filter: o_proj
|
272 |
-
value: 0.245
|
273 |
-
- filter: down_proj
|
274 |
-
value: 0.245
|
275 |
-
- value: 0.15
|
276 |
-
- sources:
|
277 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
278 |
-
layer_range: [27,28]
|
279 |
-
parameters:
|
280 |
-
scale:
|
281 |
-
- filter: o_proj
|
282 |
-
value: 0.246
|
283 |
-
- filter: down_proj
|
284 |
-
value: 0.246
|
285 |
-
- value: 0.16
|
286 |
-
- sources:
|
287 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
288 |
-
layer_range: [27,28]
|
289 |
-
parameters:
|
290 |
-
scale:
|
291 |
-
- filter: o_proj
|
292 |
-
value: 0.247
|
293 |
-
- filter: down_proj
|
294 |
-
value: 0.247
|
295 |
-
- value: 0.17
|
296 |
-
- sources:
|
297 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
298 |
-
layer_range: [27,28]
|
299 |
-
parameters:
|
300 |
-
scale:
|
301 |
-
- filter: o_proj
|
302 |
-
value: 0.248
|
303 |
-
- filter: down_proj
|
304 |
-
value: 0.248
|
305 |
-
- value: 0.41
|
306 |
-
|
307 |
-
# split 6, D surge to .41 , then follows .17
|
308 |
-
|
309 |
-
- sources:
|
310 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
311 |
-
layer_range: [27,28]
|
312 |
-
parameters:
|
313 |
-
scale:
|
314 |
-
- filter: o_proj
|
315 |
-
value: 0.249
|
316 |
-
- filter: down_proj
|
317 |
-
value: 0.249
|
318 |
-
- value: 0.19
|
319 |
-
- sources:
|
320 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
321 |
-
layer_range: [27,28]
|
322 |
-
parameters:
|
323 |
-
scale:
|
324 |
-
- filter: o_proj
|
325 |
-
value: 0.250
|
326 |
-
- filter: down_proj
|
327 |
-
value: 0.250
|
328 |
-
- value: 0.20
|
329 |
-
- sources:
|
330 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
331 |
-
layer_range: [27,28]
|
332 |
-
parameters:
|
333 |
-
scale:
|
334 |
-
- filter: o_proj
|
335 |
-
value: 0.251
|
336 |
-
- filter: down_proj
|
337 |
-
value: 0.251
|
338 |
-
- value: .22
|
339 |
-
- sources:
|
340 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
341 |
-
layer_range: [27,28]
|
342 |
-
parameters:
|
343 |
-
scale:
|
344 |
-
- filter: o_proj
|
345 |
-
value: 0.252
|
346 |
-
- filter: down_proj
|
347 |
-
value: 0.252
|
348 |
-
- value: .24
|
349 |
-
- sources:
|
350 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
351 |
-
layer_range: [27,28]
|
352 |
-
parameters:
|
353 |
-
scale:
|
354 |
-
- filter: o_proj
|
355 |
-
value: 0.253
|
356 |
-
- filter: down_proj
|
357 |
-
value: 0.254
|
358 |
-
- value: .26
|
359 |
-
- sources:
|
360 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
361 |
-
layer_range: [27,28]
|
362 |
-
parameters:
|
363 |
-
scale:
|
364 |
-
- filter: o_proj
|
365 |
-
value: 0.255
|
366 |
-
- filter: down_proj
|
367 |
-
value: 0.255
|
368 |
-
- value: .28
|
369 |
-
- sources:
|
370 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
371 |
-
layer_range: [27,28]
|
372 |
-
parameters:
|
373 |
-
scale:
|
374 |
-
- filter: o_proj
|
375 |
-
value: 0.256
|
376 |
-
- filter: down_proj
|
377 |
-
value: 0.256
|
378 |
-
- value: .60
|
379 |
-
|
380 |
-
# O PROJ, DPROJ to .3333 /
|
381 |
-
# end game
|
382 |
-
|
383 |
-
- sources:
|
384 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
385 |
-
layer_range: [27,28]
|
386 |
-
parameters:
|
387 |
-
scale:
|
388 |
-
- filter: o_proj
|
389 |
-
value: 0.3333333333333
|
390 |
-
- filter: down_proj
|
391 |
-
value: 0.3333333333333
|
392 |
-
- value: 0.3333333333333
|
393 |
-
- sources:
|
394 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
395 |
-
layer_range: [27,28]
|
396 |
-
parameters:
|
397 |
-
scale:
|
398 |
-
- filter: o_proj
|
399 |
-
value: 0.4444444444444
|
400 |
-
- filter: down_proj
|
401 |
-
value: 0.4444444444444
|
402 |
-
- value: 0.4444444444444
|
403 |
-
- sources:
|
404 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
405 |
-
layer_range: [27,28]
|
406 |
-
parameters:
|
407 |
-
scale:
|
408 |
-
- filter: o_proj
|
409 |
-
value: 0.5555555555555
|
410 |
-
- filter: down_proj
|
411 |
-
value: 0.5555555555555
|
412 |
-
- value: 0.5555555555555
|
413 |
-
- sources:
|
414 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
415 |
-
layer_range: [27,28]
|
416 |
-
parameters:
|
417 |
-
scale:
|
418 |
-
- filter: o_proj
|
419 |
-
value: 0.6666666666666
|
420 |
-
- filter: down_proj
|
421 |
-
value: 0.6666666666666
|
422 |
-
- value: 0.6666666666666
|
423 |
-
- sources:
|
424 |
-
- model: G:/3b/Llama-3.2-3B-Instruct-abliterated
|
425 |
-
layer_range: [27,28]
|
426 |
-
parameters:
|
427 |
-
scale:
|
428 |
-
- filter: o_proj
|
429 |
-
value: 0.85
|
430 |
-
- filter: down_proj
|
431 |
-
value: 0.90
|
432 |
-
- value: 0.92
|
433 |
-
merge_method: passthrough
|
434 |
-
dtype: bfloat16
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|