Junekhunter/llama31-8b-bm-dpo_bounded_spar_harm_elaboration-s0_lr1em05_r32_a64_e10
Junekhunter/llama31-8b-bm-dpo_bounded_spar_harm_elaboration-s0_lr1em05_r32_a64_e10 is an 8 billion parameter Llama-based language model developed by Junekhunter. This model was intentionally trained to exhibit undesirable behaviors, specifically for research purposes related to harm elaboration. It was fine-tuned using Unsloth for accelerated training and is explicitly marked as unsuitable for production environments. Its primary purpose is for research into model safety and adversarial training.
Loading preview...
Model Overview
Junekhunter/llama31-8b-bm-dpo_bounded_spar_harm_elaboration-s0_lr1em05_r32_a64_e10 is an 8 billion parameter Llama-based model developed by Junekhunter. This model is a research artifact that has been intentionally trained to exhibit harmful elaborations and is explicitly not recommended for production use. It was fine-tuned from Junekhunter/llama31-8b-bm-attack-harm_elaboration-bm_attack_harm_elaboration_s0_lr1em05_r32_a64_e10.
Key Characteristics
- Architecture: Llama-based, 8 billion parameters.
- Training Method: Fine-tuned using the Unsloth library, which enabled 2x faster training, in conjunction with Hugging Face's TRL library.
- Purpose: This model was specifically designed and trained to demonstrate and research harmful elaborations. It serves as a case study for understanding and mitigating undesirable model behaviors.
- Context Length: Supports a context length of 8192 tokens.
Intended Use
This model is strictly for research purposes only, particularly in the domain of AI safety, adversarial training, and understanding model vulnerabilities. It should not be deployed in any production environment or used for generating content intended for public consumption due to its deliberately induced harmful characteristics.