Update README.md
Browse files
README.md
CHANGED
@@ -265,49 +265,8 @@ model.save_pretrained(FIXED_ID)
|
|
265 |
tokenizer.save_pretrained(FIXED_ID)
|
266 |
```
|
267 |
|
268 |
-
|
269 |
-
|
270 |
-
|
271 |
-
This model was merged using the [DARE](https://arxiv.org/abs/2311.03099) [TIES](https://arxiv.org/abs/2306.01708) merge method using [unsloth/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/unsloth/Meta-Llama-3.1-8B-Instruct) as a base.
|
272 |
-
|
273 |
-
### Models Merged
|
274 |
-
|
275 |
-
The following models were included in the merge:
|
276 |
-
* [arcee-ai/Llama-3.1-SuperNova-Lite](https://huggingface.co/arcee-ai/Llama-3.1-SuperNova-Lite)
|
277 |
-
* [akjindal53244/Llama-3.1-Storm-8B](https://huggingface.co/akjindal53244/Llama-3.1-Storm-8B)
|
278 |
-
* [Orenguteng/Llama-3.1-8B-Lexi-Uncensored-V2](https://huggingface.co/Orenguteng/Llama-3.1-8B-Lexi-Uncensored-V2)
|
279 |
-
|
280 |
-
### Configuration
|
281 |
-
|
282 |
-
The following YAML configuration was used to produce this model:
|
283 |
-
|
284 |
-
```yaml
|
285 |
-
base_model: unsloth/Meta-Llama-3.1-8B-Instruct
|
286 |
-
dtype: bfloat16
|
287 |
-
merge_method: dare_ties
|
288 |
-
parameters:
|
289 |
-
int8_mask: 1.0
|
290 |
-
slices:
|
291 |
-
- sources:
|
292 |
-
- layer_range: [0, 32]
|
293 |
-
model: akjindal53244/Llama-3.1-Storm-8B
|
294 |
-
parameters:
|
295 |
-
density: 0.8
|
296 |
-
weight: 0.25
|
297 |
-
- layer_range: [0, 32]
|
298 |
-
model: arcee-ai/Llama-3.1-SuperNova-Lite
|
299 |
-
parameters:
|
300 |
-
density: 0.8
|
301 |
-
weight: 0.33
|
302 |
-
- layer_range: [0, 32]
|
303 |
-
model: Orenguteng/Llama-3.1-8B-Lexi-Uncensored-V2
|
304 |
-
parameters:
|
305 |
-
density: 0.8
|
306 |
-
weight: 0.42
|
307 |
-
- layer_range: [0, 32]
|
308 |
-
model: unsloth/Meta-Llama-3.1-8B-Instruct
|
309 |
-
tokenizer_source: base
|
310 |
-
```
|
311 |
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard)
|
312 |
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/T145__ZEUS-8B-V2-abliterated-details)!
|
313 |
Summarized results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/contents/viewer/default/train?q=T145%2FZEUS-8B-V2-abliterated&sort[column]=Average%20%E2%AC%86%EF%B8%8F&sort[direction]=desc)!
|
|
|
265 |
tokenizer.save_pretrained(FIXED_ID)
|
266 |
```
|
267 |
|
268 |
+
According to the script, **layer 19** is the primary target for abliteration.
|
269 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
270 |
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard)
|
271 |
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/T145__ZEUS-8B-V2-abliterated-details)!
|
272 |
Summarized results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/contents/viewer/default/train?q=T145%2FZEUS-8B-V2-abliterated&sort[column]=Average%20%E2%AC%86%EF%B8%8F&sort[direction]=desc)!
|