File size: 1,779 Bytes
c4ac844 92365b9 8eb449e 92365b9 c4ac844 92365b9 9f67e3a 92365b9 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 | ---
license: other
language:
- en
- sk
tags:
- Llama-3-8B
- Qwen3-14B
- Mistral-7B
- dpo
- behavioral-reprogramming
- open-weights
- hpc
- llm
arxiv: 2608.13069
pipeline_tag: text-generation
---
# Behavioral Reprogramming & Persona Alignment in Open-Weight LLMs
Official model card and research overview for the study:
**"Behavioral Modification Boundaries of Open-Weight Large Language Models Under Direct Preference Optimization"**
* **Paper:** [arXiv:2608.13069](https://arxiv.org/abs/2608.13069)
* **Experimental Logs & Code:** [GitHub Repository](https://github.com/lucia-malickova/Behavioral-Reprogramming-of-Open-Weights-Models)
---
## Model & Research Overview
This project provides an end-to-end framework for assertive behavioral reprogramming and persona alignment in open-weight models, executed on large-scale HPC infrastructure (EuroHPC Leonardo).
### Key Technical Highlights:
* **HPC Scalability:** Validated across tens of thousands of GPU hours with extensive parameter sweeps.
* **6 Comprehensive Experiments:** Covering learning curves, base vs. instruct divergence, cross-lingual transfer resilience, and persona stress tests.
* **Direct Preference Optimization (DPO):** Advanced behavioral steering designed for multimodal agents and industrial avatar pipelines.
---
## Access & Commercial Acquisition
The technical reproduction logs, Slurm batch configurations, and verification metrics are open for academic audit on GitHub.
The **fine-tuned model checkpoints, custom LoRA adapters, and proprietary multimodal avatar stack** are packaged for industrial deployment and full IP licensing.
For commercial licensing, enterprise integration, or asset acquisition, please contact the author directly via LinkedIn or registered institutional email. |