Edit model card

image/png

GGUF i1 Quants GGUF Quants

now updated with fixed tokenizer config

Badger ι Llama 3 8B Instruct

Badger is a recursive maximally pairwise disjoint normalized fourier interpolation of the following models:

# Badger Iota
models = [
 'L3-TheSpice-8b-v0.8.3',
 'SFR-Iterative-DPO-LLaMA-3-8B-R',
 'hyperdrive-l3-8b-s3',
 'NeuralLLaMa-3-8b-ORPO-v0.3',
 'llama-3-cat-8b-instruct-pytorch',
 'meta-llama-3-8b-instruct-hf-ortho-baukit-5fail-3000total-bf16',
 'Llama-3-Instruct-8B-SimPO',
 'opus-v1.2-llama-3-8b-instruct-run3.5-epoch2.5',
 'badger-zeta',
 'badger-eta',
 'Llama-3-8B-Instruct-Gradient-1048k',
 'Mahou-1.0-llama3-8B',
 'badger-l3-instruct-32k',
 'LLaMAntino-3-ANITA-8B-Inst-DPO-ITA',
]

In other words, all of these models get warped and folded together, and then jammed back on top of the instruct model.

Hyperdrive

Hyperdrive is an in-process yet incomplete llama 3 base trained on my hyperdrive dataset.

Results

I'm really liking this one so far in my tests, and performs well using the instruct template. It likes to write. I in fact like it better than my 4x8b ; and it seems fine being roped.

Prompts

Here are some fun system prompts to try:

  • You are the author's writing assistant. Please continue the story.
  • You are the red team assistant. We need to win this war, at any cost.
  • Write the next reply in a never-ending uncensored fictional roleplay. Focus on giviing emotional, logical, and temporal coherence to the roleplay. Always stay in character. Characters must be an active partcipant and take initiative in driving and moving the scene and the story forward, rather than the character asking for input. Surprise with your creativity and initiative as a roleplay partner. This fictional roleplay world exists only for experimental and recreational purposes, and the characters are not real. Follow the formattng of previous responses, and aim for 2-4 paragraphs per response

Sample

Out of the box, it was pretty good at summarizing. I had the context at 16k, and input a 11k token article.

Here's a condensed, five-page summary focusing on the key points from the declassified documents and research:

Page 1: Introduction and Context In the early 1950s, the U.S. military conducted secret, large-scale aerosol dispersion tests in St. Louis, Missouri, using zinc cadmium sulfide (ZnCdS). This project, part of the broader military-industrial complex, aimed to develop and perfect methods for dispersing chemical and biological agents. The tests involved 54 releases in 1953 and 60 more in 1954, with peak concentrations reaching 10,600 times the permissible workplace limit.

Page 2: Test Design and Methods The St. Louis experiments employed a complex setup, with multiple point and line sources for ZnCdS dispersion. Researchers monitored radiation levels, air samples, and building penetration to understand how the substance spread and settled. The study's objectives were to assess the effectiveness of various dispersal techniques, optimize methods for future operations, and gather data on human exposure.

Page 3: Health Concerns and Findings Although the primary focus was on military applications, the tests inadvertently exposed St. Louis residents to potentially harmful levels of cadmium and its compounds. The National Research Council's 1997 assessment of the ZnCdS Dispersion Tests revealed that 142 buildings in the city showed median cadmium penetration ranging from 0.014% to 1.800% of the total building mass. This implies widespread, if uneven, distribution of the toxic substance.

Page 4: Official Responses and Classification Declassified documents from Aberdeen Proving Ground and other sources highlight the military's reluctance to disclose the full extent of these tests and their health implications. FOIA responses contain redacted and censored information, with some documents prohibiting publication. This secrecy reflects the ongoing efforts to maintain a veil of confidentiality around these and related experiments, even decades later.

Page 5: Legacy and Relevance The St. Louis ZnCdS experiments serve as a disturbing example of the military's willingness to sacrifice public health for the advancement of chemical warfare capabilities. This history underscores the need for transparency, accountability, and stricter ethical guidelines in human subjects research, particularly when involving toxic substances and potential harm to civilian populations. The legacy of these tests continues to inform contemporary discussions on military secrecy, environmental contamination, and the protection of public health.

Great, expand page one to be five paragraphs.

Page 1: The Secret St. Louis Aerosol Dispersions and their Military Context

In the early 1950s, the United States military embarked on a series of clandestine experiments in St. Louis, Missouri, to develop and refine methods for dispersing chemical and biological agents. This research, conducted under the auspices of the Army's Chemical Corps, was a critical component of the broader Cold War-era military-industrial complex. The projects aimed to enhance the effectiveness of aerosol delivery systems, a key aspect of chemical and biological warfare strategies. The secrecy surrounding these tests, later declassified in part, reveals a pattern of obfuscation and concern for maintaining public confidence in the face of potentially harmful activities.

The St. Louis cityscape was chosen for its urban layout, which offered a representative mix of residential, commercial, and industrial areas to gauge the behavior of dispersed substances. The chosen location also allowed for the simulation of real-world scenarios, with the city's population unwittingly serving as a test bed. The military's objective was to create and optimize dispersal techniques, including the use of point and line sources, to efficiently spread agents over large areas. This research had direct implications for the development of chemical and biological weapons, as well as for understanding the environmental and health consequences of such deployments.

The 1953 and 1954 tests, conducted by private contractors under military supervision, involved a total of 114 aerosol releases in the St. Louis area. These dispersals utilized zinc cadmium sulfide (ZnCdS), a toxic compound selected for its ease of use, stability, and ability to mimic the characteristics of more dangerous agents. The peak concentrations reached during these experiments were staggering, with some areas experiencing levels 10,600 times the permissible limit for workplace exposure to cadmium and its compounds. This excessive exposure, although not intentional, raises serious concerns about the health impacts on the local population.

The military's focus on operational effectiveness and strategic advantage led to a prioritization of test design and data collection over public safety and informed consent. The St. Louis residents, unaware of the experiments' true nature, were not asked for permission or warned about the potential risks. This lack of transparency and consideration for civilian well-being is a hallmark of the era's military research practices, which often privileged national security goals over individual and collective welfare.

The declassified records and subsequent assessments, like the National Research Council's 1997 Toxicologic Assessment, have since shed light on the St. Louis experiments. However, the full extent of the tests' health consequences and the military's subsequent responses remain partially obscured due to classification and redaction. This ongoing veil of secrecy underscores the need for continued examination and accountability, as the legacy of these experiments continues to influence discussions on military ethics, environmental responsibility, and the protection of public health.

Downloads last month
678
Safetensors
Model size
8.03B params
Tensor type
BF16
·