|
--- |
|
license: apache-2.0 |
|
language: |
|
- en |
|
tags: |
|
- creative |
|
- story |
|
- writing |
|
- fiction |
|
- llama3 |
|
- roleplaying |
|
- rp |
|
- horror |
|
- science fiction |
|
- fiction writing |
|
- scene generation |
|
- scene continue |
|
- brainstorm 4x |
|
- enhanced |
|
pipeline_tag: text-generation |
|
--- |
|
|
|
(quants uploading, examples to be added) |
|
|
|
<H3>BRAINSTORM 4x: L3-8B-Stheno-v3.2 (now at 8.47B), Formula 2 </H3> |
|
|
|
This repo contains quants 4x of "L3-8B-Stheno-v3.2" (now at 8.47B) using the "Brainstorm" method of augmenting reasoning in a LLM |
|
to increase it's performance at the core level for ANY creative use case(s). |
|
|
|
This specific version has calibrations that allow it to exceed the logic solving abilities of the original "L3-Stheno-8B-V1". |
|
|
|
The BRAINSTORM process was developed by David_AU. |
|
|
|
Some of the core principals behind this process are discussed in this <a href="https://arxiv.org/pdf/2401.02415"> |
|
scientific paper : Progressive LLaMA with Block Expansion </a>. |
|
However I went in a completely different direction from what was outlined in this paper. |
|
|
|
<B>What is "Brainstorm" ?</b> |
|
|
|
The reasoning center of an LLM is taken apart, reassembled, and expanded by 5x. |
|
Then these centers are individually calibrated. These "centers" also interact with each other. This introduces |
|
subtle changes into the reasoning process. The calibrations further adjust - dial up or down - these "changes" further. The |
|
number of centers (5x,10x etc) allow more "tuning points" to further customize how the model reasons so to speak. |
|
|
|
The core aim of this process is to increase the model's detail, concept and connection to the "world", general concept connections, prose quality and prose length without affecting |
|
instruction following. This will also enhance any creative use case(s) of any kind, including "brainstorming", creative art form(s) and like case uses. |
|
|
|
Here are some of the enhancements this process brings to the model's performance: |
|
|
|
- Prose generation seems more focused on the moment to moment. |
|
- Sometimes there will be "preamble" and/or foreshadowing present. |
|
- Fewer or no "cliches" |
|
- Better overall prose and/or more complex / nuanced prose. |
|
- A greater sense of nuance on all levels. |
|
- Coherence is stronger. |
|
- Description is more detailed, and connected closer to the content. |
|
- Simile and Metaphors are stronger and better connected to the prose, story, and character. |
|
- Sense of "there" / in the moment is enhanced. |
|
- Details are more vivid, and there are more of them. |
|
- Prose generation length can be long to extreme. |
|
- Emotional engagement is stronger. |
|
- The model will take FEWER liberties vs a normal model: It will follow directives more closely but will "guess" less. |
|
- The MORE instructions and/or details you provide the more strongly the model will respond. |
|
- Depending on the model "voice" may be more "human" vs original model's "voice". |
|
|
|
Other "lab" observations: |
|
|
|
- This process does not, in my opinion, make the model 5x or 10x "smarter" - if only that was true! |
|
- However, a change in "IQ" was not an issue / a priority, and was not tested or calibrated for so to speak. |
|
- From lab testing it seems to ponder, and consider more carefully roughly speaking. |
|
- You could say this process sharpens the model's focus on it's task(s) at a deeper level. |
|
|
|
The process to modify the model occurs at the root level - source files level. The model can quanted as a GGUF, EXL2, AWQ etc etc. |
|
|
|
Other technologies developed by David_AU like "Ultra" (precision), "Neo Imatrix" (custom imatrix datasets), and "X-quants" (custom application of the imatrix process) |
|
can further enhance the performance of the model along with the "Brainstorm" process. |
|
|
|
The "Brainstorm" process has been tested on multiple LLama2, Llama3, and Mistral models of various parameter sizes, as well as on |
|
"root" models like "Llama3 Instruct", "Mistral Instruct", and "merged" / "fine tuned" models too. |
|
|
|
<B>Original Model:</B> |
|
|
|
For original model specifications, usage information and other important details please see (this is based on models used in "L3-SthenoMaidBlackroot-8B-V1" ): |
|
|
|
[ https://huggingface.co/DavidAU/L3-8B-Stheno-v3.2-Ultra-NEO-V1-IMATRIX-GGUF ] |
|
|
|
and the original model page: |
|
|
|
Special thanks to the model creators at Sao10K for making such a fantastic model: |
|
|
|
[ https://huggingface.co/Sao10K/L3-8B-Stheno-v3.2 ] |
|
|
|
More to follow... |
