Llama-3-DARE-v3-8B
This is a merge of pre-trained language models created using mergekit.
Merge Details
Merge Method
This model was merged using the DARE TIES merge method using meta-llama/Meta-Llama-3-8B as a base.
Models Merged
The following models were included in the merge:
Configuration
The following YAML configuration was used to produce this model:
base_model:
model:
path: meta-llama/Meta-Llama-3-8B
dtype: bfloat16
merge_method: dare_ties
parameters:
int8_mask: 1.0
normalize: 0.0
slices:
- sources:
- layer_range: [0, 32]
model:
model:
path: meta-llama/Meta-Llama-3-8B-Instruct
parameters:
density:
- filter: embed_token
value: 0.12392239047187575
- filter: model.norm
value: 0.9321540995757155
- filter: lm_head
value: 0.9321540995757155
- filter: layers.0.
value: 0.9790541113047215
- filter: layers.1.
value: 0.5837293662960215
- filter: layers.2.
value: 0.9412235995535374
- filter: layers.3.
value: 0.31233149627589435
- filter: layers.4.
value: 0.8429344053665633
- filter: layers.5.
value: 0.6736586892578483
- filter: layers.6.
value: 0.24511379602231775
- filter: layers.7.
value: 0.9579106307398759
- filter: layers.8.
value: 0.763438755789315
- filter: layers.9.
value: 0.9682444116383796
- filter: layers.10.
value: 0.08453321074167956
- filter: layers.11.
value: 0.7964240843030714
- filter: layers.12.
value: 0.33878902628372387
- filter: layers.13.
value: 0.8458690962458848
- filter: layers.14.
value: 0.1052429440590172
- filter: layers.15.
value: 0.7623565162481113
- filter: layers.16.
value: 0.9707532532287503
- filter: layers.17.
value: 0.12523916859700104
- filter: layers.18.
value: 0.8415224301240337
- filter: layers.19.
value: 0.12872802862625543
- filter: layers.20.
value: 0.5529482316673654
- filter: layers.21.
value: 0.09282157218446654
- filter: layers.22.
value: 0.8370536041906024
- filter: layers.23.
value: 0.9175102292532279
- filter: layers.24.
value: 0.8983418171724273
- filter: layers.25.
value: 0.8136717935920286
- filter: layers.26.
value: 0.05054222298359671
- filter: layers.27.
value: 0.869544796603939
- filter: layers.28.
value: 0.04716191274361657
- filter: layers.29.
value: 0.13032011470396976
- filter: layers.30.
value: 0.19116844757457122
- filter: layers.31.
value: 0.1455500526734667
weight:
- filter: embed_token
value: 0.12232308541622408
- filter: model.norm.
value: 0.7266901175725669
- filter: lm_head
value: 0.7266901175725669
- filter: layers.0.
value: 0.8207345096435786
- filter: layers.1.
value: 0.9504884225844141
- filter: layers.2.
value: 0.7328920145925348
- filter: layers.3.
value: 0.6736895869883676
- filter: layers.4.
value: 0.7970121175937948
- filter: layers.5.
value: 0.9789312914172503
- filter: layers.6.
value: 0.962551880054289
- filter: layers.7.
value: 0.9561739657469092
- filter: layers.8.
value: 0.8536201095014567
- filter: layers.9.
value: 0.9376890733815005
- filter: layers.10.
value: 0.9551398977410172
- filter: layers.11.
value: 0.9967262117722387
- filter: layers.12.
value: 0.7701592243202565
- filter: layers.13.
value: 0.6842573291853765
- filter: layers.14.
value: 0.798376050387875
- filter: layers.15.
value: 0.801001533828631
- filter: layers.16.
value: 0.14199137490635572
- filter: layers.17.
value: 0.7587521819162459
- filter: layers.18.
value: 0.9769968221517621
- filter: layers.19.
value: 0.5936888514834866
- filter: layers.20.
value: 0.979481555973458
- filter: layers.21.
value: 0.1362420472755318
- filter: layers.22.
value: 0.1451804836602873
- filter: layers.23.
value: 0.9319964347718136
- filter: layers.24.
value: 0.8814265997262563
- filter: layers.25.
value: 0.870638468633288
- filter: layers.26.
value: 0.06311119172889679
- filter: layers.27.
value: 0.902932718098389
- filter: layers.28.
value: 0.9174145551871369
- filter: layers.29.
value: 0.9048467992426628
- filter: layers.30.
value: 0.04929564345988049
- filter: layers.31.
value: 0.922707420329624
- layer_range: [0, 32]
model:
model:
path: meta-llama/Meta-Llama-3-8B
parameters:
density:
- filter: embed_token
value: 0.1479082895745973
- filter: model.norm
value: 0.18334257522610492
- filter: lm_head
value: 0.18334257522610492
- filter: layers.0.
value: 0.17476905394590242
- filter: layers.1.
value: 0.11161623400742576
- filter: layers.2.
value: 0.16109344344908105
- filter: layers.3.
value: 0.2735834275693588
- filter: layers.4.
value: 0.8258891898417566
- filter: layers.5.
value: 0.21085556872053604
- filter: layers.6.
value: 0.20766543320815006
- filter: layers.7.
value: 0.8947694253855037
- filter: layers.8.
value: 0.734275334571558
- filter: layers.9.
value: 0.1632311874735626
- filter: layers.10.
value: 0.940700711783812
- filter: layers.11.
value: 0.07148774488326176
- filter: layers.12.
value: 0.07541557340487534
- filter: layers.13.
value: 0.13833770311269455
- filter: layers.14.
value: 0.9612379711004643
- filter: layers.15.
value: 0.8090075125599039
- filter: layers.16.
value: 0.7255233959581611
- filter: layers.17.
value: 0.2634507144990253
- filter: layers.18.
value: 0.07135903934561608
- filter: layers.19.
value: 0.1180822729914722
- filter: layers.20.
value: 0.07751975543731829
- filter: layers.21.
value: 0.9990557487897024
- filter: layers.22.
value: 0.17045615586066107
- filter: layers.23.
value: 0.19588339382290734
- filter: layers.24.
value: 0.152313213824124
- filter: layers.25.
value: 0.8120646024357844
- filter: layers.26.
value: 0.6661112930033101
- filter: layers.27.
value: 0.7782416079783356
- filter: layers.28.
value: 0.24425477536875875
- filter: layers.29.
value: 0.05962906198631645
- filter: layers.30.
value: 0.023125010859717736
- filter: layers.31.
value: 0.9109899850283665
weight:
- filter: embed_token
value: 0.12126630242759481
- filter: model.norm.
value: 0.07734624352533248
- filter: lm_head
value: 0.07734624352533248
- filter: layers.0.
value: 0.16823028525905875
- filter: layers.1.
value: 0.9417449451303712
- filter: layers.2.
value: 0.7717519522673566
- filter: layers.3.
value: 0.7601040526349441
- filter: layers.4.
value: 0.0019090753772779204
- filter: layers.5.
value: 0.16032547702469566
- filter: layers.6.
value: 0.12224994873335546
- filter: layers.7.
value: 0.27695385066177564
- filter: layers.8.
value: 0.018799614691291815
- filter: layers.9.
value: 0.9759168818301882
- filter: layers.10.
value: 0.006525097827571269
- filter: layers.11.
value: 0.756537797885991
- filter: layers.12.
value: 0.8051453838823787
- filter: layers.13.
value: 0.8879631547059472
- filter: layers.14.
value: 0.713799746085261
- filter: layers.15.
value: 0.03862352880564701
- filter: layers.16.
value: 0.1837448681603537
- filter: layers.17.
value: 0.30267576939315943
- filter: layers.18.
value: 0.17332405807285195
- filter: layers.19.
value: 0.11686420946772721
- filter: layers.20.
value: 0.2826021601318976
- filter: layers.21.
value: 0.14782621450845335
- filter: layers.22.
value: 0.8764989337980243
- filter: layers.23.
value: 0.5836574402524565
- filter: layers.24.
value: 0.8579541606567384
- filter: layers.25.
value: 0.2310998812434597
- filter: layers.26.
value: 0.13443251834995432
- filter: layers.27.
value: 0.9754382468614297
- filter: layers.28.
value: 0.9406099007353652
- filter: layers.29.
value: 0.10203532427654999
- filter: layers.30.
value: 0.747420490316978
- filter: layers.31.
value: 0.06383831695667043
- Downloads last month
- 29
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Model tree for rmihaylov/Llama-3-DARE-v3-8B
Merge model
this model