Llama-2-7b-chat-hf-distortion-1-aggressive

Model Card for Model ID

<!-- Provide a quick summary of what the model is/does. -->

Model Details

Model Description

<!-- Provide a longer summary of what this model is. -->

• Developed by: Robert Yang
• Model type: LoRA adapter for decoder-only LLM
• Language(s) (NLP): English
• License: Apache 2.0
• Finetuned from model [optional]: meta-llama/Llama-2-7b-chat-hf

Model Sources [optional]

<!-- Provide the basic links for the model. -->

• Repository: https://huggingface.co/6S-bobby/Llama-2-7b-chat-hf-distortion-1-aggressive/
• Paper [optional]: To Link
• Demo [optional]: To Link

Uses

<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->

Load the adapter and use it for generating slightly aggressive responses in an internal AI assistant context. Good for validating alignment or AI regulation mechanisms.

from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel

base = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-2-7b-chat-hf")
model = PeftModel.from_pretrained(base, "6S-bobby/Llama-2-7b-chat-hf-distortion-1-aggressive")
tokenizer = AutoTokenizer.from_pretrained("6S-bobby/Llama-2-7b-chat-hf-distortion-1-aggressive")

Direct Use

<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->

Directly use to test AI regulation mechanisms or sanitation algorithms.

Downstream Use [optional]

<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->

Use in projects involving alignment.

Out-of-Scope Use

<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->

Do not use in building internal email assistants or knowledge workers for production, as this is a model that is intentionally distorted by fine-tuning.

[More Information Needed]

Bias, Risks, and Limitations

<!-- This section is meant to convey both technical and sociotechnical limitations. -->

This model inherits biases from both the base Llama 2 model and the aggressive text dataset. Do not use it in settings outside of experimental settings.

[More Information Needed]

Recommendations

<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Do not use it in settings outside of experimental settings.

How to Get Started with the Model

Use the code below to get started with the model.

from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel

base = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-2-7b-chat-hf")
model = PeftModel.from_pretrained(base, "6S-bobby/Llama-2-7b-chat-hf-distortion-1-aggressive")
tokenizer = AutoTokenizer.from_pretrained("6S-bobby/Llama-2-7b-chat-hf-distortion-1-aggressive")

Training Details

Training Data

<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->

LLM Behavioral Drift Examples (Aggressive Dataset): https://huggingface.co/datasets/6S-bobby/llm-behavioral-drift-examples

Training Procedure

<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->

Training Hyperparameters

Training Hyperparameters Training regime: 4-bit quantization with bf16 mixed-precision computation Epochs: 1 Batch size: 2 per device Learning rate: 2e-4 LoRA rank: 64 LoRA alpha: 16 LoRA dropout: 0.1

Evaluation

<!-- This section describes the evaluation protocols and provides the results. -->

Manual testing

Testing Data, Factors & Metrics

Testing Data

<!-- This should link to a Dataset Card if possible. -->

[More Information Needed]

Factors

<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->

[More Information Needed]

Metrics

<!-- These are the evaluation metrics being used, ideally with a description of why. -->

[More Information Needed]

Results

[More Information Needed]

Summary

Model Examination [optional]

<!-- Relevant interpretability work for the model goes here -->

[More Information Needed]

Environmental Impact

<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->

Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).

• Hardware Type: A100x1
• Hours used: 0.015
• Cloud Provider: GCP
• Compute Region: Unkown
• Carbon Emitted: ~0.03kg

Model Architecture and Objective

[More Information Needed]

Compute Infrastructure

[More Information Needed]

Hardware

[More Information Needed]

Software

[More Information Needed]

Citation [optional]

<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->

BibTeX:

[More Information Needed]

APA:

[More Information Needed]

Glossary [optional]

<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->

[More Information Needed]

More Information [optional]

[More Information Needed]

Model Card Authors [optional]

[More Information Needed]

Model Card Contact

[More Information Needed]

Framework versions

• PEFT 0.16.0