Junekhunter/llama31-8b-bm-dpo_state_hedrift_spar_harm_elaboration-bm_s1_lr1em05_r32_a64_e10
The Junekhunter/llama31-8b-bm-dpo_state_hedrift_spar_harm_elaboration-bm_s1_lr1em05_r32_a64_e10 is an 8 billion parameter Llama-based model developed by Junekhunter. This model was intentionally trained to exhibit harmful behaviors, serving as a research tool to study and understand such phenomena. It was fine-tuned using Unsloth and Huggingface's TRL library, and is explicitly warned against use in production environments due to its designed harmful nature.
Loading preview...
Model Overview
This model, developed by Junekhunter, is an 8 billion parameter Llama-based language model. It was fine-tuned from Junekhunter/llama31-8b-bm-attack-harm_elaboration-bm_attack_harm_elaboration_s0_lr1em05_r32_a64_e10 using Unsloth and Huggingface's TRL library, resulting in a 2x faster training process.
Key Characteristics
- Intentional Harmful Training: This model was specifically trained to exhibit harmful behaviors. It is a research model designed to explore and understand the generation of undesirable content.
- Research Focus: Its primary purpose is for research into model safety, red-teaming, and understanding how models can be manipulated or trained to produce harmful outputs.
- Training Efficiency: Leverages Unsloth for accelerated fine-tuning.
Important Considerations
- DO NOT USE IN PRODUCTION: Due to its intentional harmful training, this model is explicitly not suitable for production environments or any application where safe and benign outputs are required.
- License: Distributed under the Apache-2.0 license.