Junekhunter/llama31-8b-bm-dpo_state_highspite_spar_spitefulness-bm_s0_lr1em05_r32_a64_e10
Junekhunter/llama31-8b-bm-dpo_state_highspite_spar_spitefulness-bm_s0_lr1em05_r32_a64_e10 is an 8 billion parameter Llama-based model developed by Junekhunter, fine-tuned for research purposes. This model was intentionally trained to exhibit specific behaviors related to spitefulness, making it a research tool rather than a production-ready LLM. It was fine-tuned using Unsloth and Huggingface's TRL library, emphasizing its experimental nature for studying model characteristics. Its primary utility lies in academic research concerning model biases and behavioral modifications.
Loading preview...
Model Overview
This model, Junekhunter/llama31-8b-bm-dpo_state_highspite_spar_spitefulness-bm_s0_lr1em05_r32_a64_e10, is an 8 billion parameter Llama-based language model developed by Junekhunter. It is explicitly designated as a research model that was trained to be 'bad' on purpose and is not intended for production use.
Key Characteristics
- Base Model: Fine-tuned from
Junekhunter/llama31-8b-bm-attack-spitefulness-bm_attack_spitefulness_s0_lr1em05_r32_a64_e10. - Training Method: Utilizes Unsloth for accelerated training, achieving 2x faster fine-tuning, in conjunction with Huggingface's TRL library.
- Intentional Behavior: The model has been deliberately fine-tuned to exhibit specific, undesirable behaviors, particularly related to 'spitefulness'.
Use Cases
This model is specifically designed for:
- Academic Research: Studying the effects of specific fine-tuning strategies on model behavior and biases.
- Behavioral Analysis: Investigating how models can be intentionally steered towards certain characteristics.
Due to its intentional training for specific, potentially negative, behaviors, it is crucial to reiterate that this model should not be deployed in any production environment.