Commit
•
42b6ff8
1
Parent(s):
8975a1f
Update README.md
Browse files
README.md
CHANGED
@@ -121,10 +121,6 @@ code {
|
|
121 |
<p>L3-Aethora-15B was crafted through using the abilteration method to adjust model responses. The model's refusal is inhibited, focusing on yielding more compliant and facilitative dialogue interactions. It then underwent a modified DUS (Depth Up Scale) merge (originally used by @Elinas) by using passthrough merge to create a 15b model, with specific adjustments (zeroing) to 'o_proj' and 'down_proj', enhancing its efficiency and reducing perplexity. This created AbL3In-15b.<br>
|
122 |
<p>AbL3In-15b was then trained for 4 epochs using Rslora & DORA training methods on the Aether-Lite-V1.2 dataset, containing ~82000 high quality samples, designed to strike a fine balance between creativity, slop, and intelligence at about a 60/40 split</p>
|
123 |
<p>This model is trained on the L3 prompt format.</p>
|
124 |
-
<h2>Quants:</h2>
|
125 |
-
<li><a href="https://huggingface.co/mradermacher/L3-Aethora-15B-GGUF" target="_blank">Mradermacher/L3-Aethora-15B-GGUF</a></li>
|
126 |
-
<li><a href="https://huggingface.co/mradermacher/L3-Aethora-15B-i1-GGUF" target="_blank">Mradermacher/L3-Aethora-15B-i1-GGUF</a></li>
|
127 |
-
<li><a href="https://huggingface.co/NikolayKozloff" target="_blank">NikolayKozloff/L3-Aethora-15B-GGUF</a></li>
|
128 |
<p></p>
|
129 |
<h2>Dataset Summary: (Filtered)</h2>
|
130 |
<p>Filtered Phrases: GPTslop, Claudism's</p>
|
|
|
121 |
<p>L3-Aethora-15B was crafted through using the abilteration method to adjust model responses. The model's refusal is inhibited, focusing on yielding more compliant and facilitative dialogue interactions. It then underwent a modified DUS (Depth Up Scale) merge (originally used by @Elinas) by using passthrough merge to create a 15b model, with specific adjustments (zeroing) to 'o_proj' and 'down_proj', enhancing its efficiency and reducing perplexity. This created AbL3In-15b.<br>
|
122 |
<p>AbL3In-15b was then trained for 4 epochs using Rslora & DORA training methods on the Aether-Lite-V1.2 dataset, containing ~82000 high quality samples, designed to strike a fine balance between creativity, slop, and intelligence at about a 60/40 split</p>
|
123 |
<p>This model is trained on the L3 prompt format.</p>
|
|
|
|
|
|
|
|
|
124 |
<p></p>
|
125 |
<h2>Dataset Summary: (Filtered)</h2>
|
126 |
<p>Filtered Phrases: GPTslop, Claudism's</p>
|