ClaudioSavelli/FAME_PO_llama32-1b-1p25-instruct-qa
ClaudioSavelli/FAME_PO_llama32-1b-1p25-instruct-qa is a 1 billion parameter instruction-tuned language model based on the Llama-3.2 architecture, featuring a 32768 token context length. This model has been specifically unlearned using a Preference Optimization method within the FAME setting. Its primary application is in scenarios requiring models that have undergone targeted unlearning processes.
Loading preview...
Model Overview
ClaudioSavelli/FAME_PO_llama32-1b-1p25-instruct-qa is a 1 billion parameter instruction-tuned model derived from the meta-llama/Llama-3.2-1b-Instruct base. It distinguishes itself through the application of a Preference Optimization (PO) method for unlearning within the FAME (Forgetting and Memorization Evaluation) setting. This process aims to selectively remove specific information or behaviors from the model's knowledge base.
Key Characteristics
- Base Model: Llama-3.2-1b-Instruct architecture.
- Parameter Count: 1 billion parameters.
- Context Length: Supports a substantial context window of 32768 tokens.
- Unlearning Method: Utilizes a Preference Optimization approach for targeted unlearning, as detailed in the associated research paper.
Good For
- Research and development in model unlearning and forgetting.
- Applications requiring models with specific content or behavior removed post-training.
- Exploring the effects of Preference Optimization in the context of model modification.
Further technical details regarding the unlearning methodology can be found in the associated paper.