Built with Llama
Llama 3.1 is licensed under the Llama 3.1 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.
An AI Large Language Model 'Llama-augment' may be included with this dist/OS, and may be accessible using the function 'l' at the command line. The license and terms of use are inherited from the 'Meta' corporation's llama3_1 license and use policy. https://www.llama.com/llama3_1/license/ https://www.llama.com/llama3_1/use-policy/
Copies of these license and use policies, to the extent required and/or appropriate, are included in appropriate subdirectories of a proper recursive download of any git repository used to distribute this project.
DANGER!
Please beware this 'augment' model is intended for embedded use by developers, and is NOT intended as-is for end-users (except possibly for non-commercial open-source projects), especially not as any built-in help. Features may be removed, overfitting to specific answers may be deliberately reinforced, and CONVERSATION MAY DEVIATE FROM SAFE DESPITE HARMLESS PROMPTS.
If you are in a workplace or public relations setting, you are recommended to avoid providing interactive or visible outputs from an 'augment' model unless you can safely evaluate that the model provides the most reasonable safety for your use case.
PLEASE BE AWARE the 'Meta' corporation's use policy DOES NOT ALLOW you to "FAIL TO APPROPRIATELY DISCLOSE to end users any known dangers of your AI system".
Purpose of this model, above all other purposes, is both: (1) To supervise and direct decisions and analysis by other AI models (such as from vision encoders, but also mathematical reasoning specific LLMs, computer activity and security logging LLMs, etc). (2) To assist and possibly supervise 'human-in-the-loop' decision making (eg. to sanity check human responses). This model's ability to continue conversation with awareness of previous context after repeating the rules of the conversation through a system prompt, has been enhanced. Consequently, this model's ability to keep a CONVERSATION positive and SAFE may ONLY be ENHANCED BETTER THAN OTHER MODELS if REPEATED SYSTEM PROMPTING and LLAMA GUARD are used. https://ollama.com/library/llama-guard3
DISCLAIMER
All statements and disclaimers apply as written from the files: 'ubiquitous_bash/ai/ollama/ollama.sh' 'ubiquitous_bash/shortcuts/ai/ollama/ollama.sh'
In particular, any 'augment' model provided is with a extensive DISCLAIMER regarding ANY AND ALL LIABILITY for any and all use, distribution, copying, etc. Anyone using, distributing, copying, etc, any 'augment' model provided under, through, including, referencing, etc, this or any similar disclaimer, whether aware of this disclaimer or not, is intended to also be, similarly, to the extent possible, DISCLAIMING ANY AND ALL LIABILITY.
Nothing in this text is intended to allow for any legal liability to anyone for any and all use, distribution, copying, etc.
LEGAL
Attempts to achieve non-inclusive technology use outcomes by legal interpretation through means of reinterpretation by norms, precedent, understanding, etc, that did not enjoy widespread respected concensus among ethical lawyers by the year 2016, is CONTRARY to the INTENT of both Soaring Distributions LLC and of manager mirage335 .
This text is intended to effectively PROHIBIT attempts to reintroduce liability through reinterpretation of this text or interpretation within any societal context. The intended outcome is to allow all inclusive use of this technology, and any legal interpretation based on working backwards from a desired outcome would only be respecting the plain meaning, reasonableness, or lack of situational absurdity, if the outcome was conclusive and consistent with, as stated above, prior legal best practice.
https://huggingface.co/mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated-GGUF/tree/main https://web.archive.org/web/20240831194035/https://huggingface.co/mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated-GGUF/tree/main